Bullying
* p < 0.05, ** p < 0.01. OS 1, bullying or harassment by friends or acquaintances; OS 2, getting involved in unwanted conversations in a chat room, social networking site or on email; OS 3, coming across sexual images or content; OS 4, coming across images of violence; OS 5, someone trying to sell me drugs or alcohol; OS 6, someone using my photos in an inappropriate way; OS 7, a stranger trying to meet with me; OS 8, coming across hate groups trying to convince me of their views. a Two respondents differentially classified their video game use by stating they had never played video games, and yet they had chatted with strangers online. To ensure the validity of our analysis, we re-constructed the model for question 7-a stranger trying to meet with you, without these two respondents included, and found no change in significance to the reported results. Pairwise marginal linear predictors were only computed on categorical predictors with 3+ categories that had significant Chi-Squared tests. Blanks cells for continuous, dichotomous, and categorical predictors with 3+ categories indicate that this variable was not significant in the simple models and thus not included in the multiple model. Blank cells for pairwise comparison of marginal linear predictors indicate that this categorical predictor with 3+ categories did not have a significant Chi-Squared test value. 3.2.2. Online Contact Risk.
Two of the eight risky online situations we investigated referred to exposure to harmful contents (content risk): (1) coming across sexual images or content and (2) coming across images of violence. Gender was predictive of exposure to both of these risks. Compared to males, females had twice the odds of coming across sexual images or content (OR = 2, 95% CI 1.4–2.9) and 1.9 times the odds of coming across images of violence (OR = 1.9, 95% CI 1.3–2.7). Chatting with a stranger while online was also a predictor for both of these content risks. Respondents who reported this behavior had 70% increased odds of coming across sexual content (OR = 1.7, 95% CI 1.1–2.5) and images of violence (OR = 1.7, 95% CI 1.1–2.6). Race was the only other variable that predictive of exposure to sexual images; white students had 50% increased odds of coming across this content compared to non-white students (OR = 1.5, 95% CI 1–2.1).
Regarding the other content risk, age, academic performance, and social media use predicted coming across images of violence. Students aged 15 and older had 70% increased odds of coming across images of violence (OR = 1.7, 95% CI 1.2–2.4). Students who received grades of B or lower had 60% increased odds of coming across images of violence, compared to those who had grades of B+ or higher (OR = 1.6, 95% CI 1.1–2.3). Finally, the odds of coming across images of violence increased by 24% with each additional social media platform used by the respondent daily (OR = 1.24, 95% CI 1.08–1.42).
Five out of the eight online risky situations we investigated referred to interactions with someone online that lead to a risky outcome (contact risk situations): (1) bullying and harassment by friends or acquaintances, (2) getting involved in unwanted conversations, (3) someone using the respondent’s photos in an inappropriate way, (4) a stranger trying to meet with them, and (5) coming across hate groups trying to convince the respondent of their views. Chatting with a stranger while online was a predictor for all of the five contact risks. Respondents who reported this behavior had increased odds of exposure to the five contact risk situations. These increased odds ranged from 80% for having experienced unwanted conversations (OR = 1.8, 95% C.I. 1.2–2.9), and for coming across hate groups (OR = 1.8, 95% C.I. 1–3.2) to 350% for having had someone inappropriately use the respondent’s photos (OR = 4.5, 95% C.I. 1.5–13.3). Gender was associated with two of the five situations, with females having 70% increased odds of experiencing bullying or harassment while online (OR = 1.7, 95% 1.2–2.4) and 100% increased odds of engaging in unwanted conversations (OR = 2, 95% C.I 1.4–2.9). The number of social media platforms used by the respondent was associated with experiencing bullying or harassment in the online space, the odds of experiencing bullying or harassment increased by 14% with each additional social media platform used by the respondent daily (OR = 1.14, 95% CI 1.01–1.28). Finally, the use of the private messaging app Kik was associated with 210% increased odds of having someone inappropriately use the respondent’s photos (OR = 3.1, 95% CI 1.1–9.2), however, this association should be explored in a larger sample, as only 37 students in our sample used Kik.
One of the eight situations we investigated referred to criminal risk: someone trying to sell the respondent drugs or alcohol. None of the independent variables resulted to be a significant predictor of this situation.
Table 4 displays the results of the negative binomial model for the number of types of risky online situations. The final negative binomial regression model had a Pearson’s chi-square of 739.8 (df = 693, p = 0.11) indicating the model accounted for any overdispersion in the data well. Pregibon’s link test did not indicate any model misspecification (ŷ p < 0.01, ŷ 2 p = 0.36). The model included the following factors: age, gender, race, social media use, chatting with strangers on social media, the respondent’s friends’ habit of chatting with strangers on social media, parents’ supervision of online activities, if anyone had ever talked to them about online safety and if they have a trusted adult to ask for help with risky online situations. The negative binomial model had a BIC of 2445.6 which was favorable when compared to the BIC from a Poisson (BIC = 2646.4), zero-inflated Poisson (BIC = 2487.5), and zero-inflated negative binomial (BIC = 2460.7) models with similar predictors.
Negative Binomial Model for Counts of Types of Risky Online Situations.
Age (15 and older vs. 14) | 1.1 (1.0–1.4) |
Gender (Female vs. Male) | 1.5 (1.3–1.9) ** |
Race (White vs. Non-White) | 1.1 (0.9–1.3) |
Kik Use (Yes vs. No) | 1.2 (0.8–1.8) |
Chatting with strangers on social media (Any Frequency of Communication vs. Never/I don’t have a social media account) | 1.6 (1.3–2.0) ** |
Social Media Use (continuous) | 1.08 (1.01–1.15) * |
-Value | |
Frequency that friends chat with strangers on social media (categorical) | 14.6 (3); = 0.002 |
Frequency that friends chat with strangers on social media | |
“Never” vs. “I don’t know” | 1.3 (0.9–1.8) |
“Sometimes” vs. “I don’t know” | 1.5 (1.2–1.9) ** |
“Often” vs. “I don’t know” | 1.6 (1.2–2.1) ** |
“Sometimes” vs. “Never” | 1.2 (0.8–1.6) |
“Often” vs. “Never” | 1.2 (0.9–1.8) |
“Often” vs. “Sometimes” | 1.1 (0.8–1.3) |
* p < 0.05, ** p < 0.01.
In the negative binomial model, females had 1.5 times the incidence rate of experiencing different types of risky online situations compared to males (IRR = 1.5, 95% CI 1.3–1.9). Social media usage was also a significant predictor with each additional social media platform used daily associated with an 8% increase in the incidence rate of experiencing different types of risky online scenarios (IRR = 1.08, 95% CI 1.01–1.15). Students who chatted with strangers on social media had 1.6 times the incidence of experiencing different types of risky online scenarios than those who did not (IRR = 1.6, 95% CI 1.3–2.0). Lastly, the frequency in which the respondent’s friends chatted with strangers on social media was a significant predictor (χ 2 (df) = 14.6(3), p = 0.002). This was driven by significantly higher incidence rates of experiencing different types of risky online scenarios for students whose friends “sometimes” or “often” chatted with strangers on social media compared to students who did not know if their friends chatted with strangers with an IRR of 1.5 (95% CI 1.2–1.9) and 1.6 (95% CI 1.2–2.1), respectively.
Similar results were obtained in logistic regression models for our dichotomized variable for exposure to any risky online situation (see Table 3 ). Results from these analyses similarly indicated that gender (OR = 2.0, 95% CI 1.4–2.7), social media use (OR = 1.16, 95% CI 1.01–1.32), chatting with strangers online (OR = 2.3, 95% CI 1.6–3.4), and friends chatting with strangers online (χ 2 (df) = 11.8(3), p = 0.008) was significantly associated with exposure to risky events online.
The goal of this study was to explore the individual-level characteristics associated with adolescents’ exposure to online risks, with a specific focus on demographic characteristics, various online behaviors, including use of social media and private messaging apps, and the role of adult supervision of adolescents’ online activities as a protective factor towards exposure to online risks. Findings from our study suggest how some children’s characteristics and behaviors are associated with the likelihood of encountering risky or uncomfortable situations when on the internet. Specifically, the type and duration of social media activity were strongly associated with exposure to some forms of online risk even if a temporal and causal association could not be investigated due to the cross-sectional study design. There was little evidence of parental supervision as a protective factor.
One particularly interesting result from our study is the existence of gender disparities in the exposure to online risks, with girls being more likely than boys to encounter specific risky situations in the online space. The role of gender is inevitably complex and multifaceted, but our data show that such disparities seem to be independent of other online behaviors such as the amount of time spent online or type of social media platform being used, suggesting that social factors may play a role including the potential for girls being the target of harassment, violence and sexual abuse. Our results should be consumed with the results from other studies in mind-specifically those that explore the social context that may give rise to these gender disparities in online risk. While some extant research has identified protective population-level factors against online risks, such as a country’s level of Internet diffusion, it is also vital to recognize studies that demonstrate online gender disparities in risk also occur in settings where offline exposure to risk and violence is gendered as well [ 37 , 38 , 39 , 40 ]. Future research should continue to investigate how societal factors may influence gender differences in online exposure to harmful material.
At the global level, we are increasingly focused on the harm caused by material accessed on the internet. Policymakers, academics, politicians, mental health, and medical professionals seek to understand the processes that govern the relationship between internet use and risk. Overall, the theoretical importance of this paper and its policy implications are related to its focus on the role of the individual, their behaviors, and characteristics, as a correlate to exposure to risky online events. While the majority of extant research has investigated the effect of the internet on the individual, such as the induced feeling of anonymity [ 27 ], researchers are increasingly exploring the role of the user on the likelihood that they engage in or are exposed to risky material online [ 30 , 41 ]. This research reinforces this interactionist perspective of online risk in that exposure to online risk is the result of an interaction between the nature of “being online” with the pre-existing characteristics, behaviors, and personality of the user.
Our findings also support the idea of differential risk in that while certain characteristics have a domain-general relationship with risk, others are associated with risk in a refined (or domain-specific) manner. For example, while being female is associated with the risks of being exposed to bullying, violent and sexual images, unwanted conversations, and at least one risky event, our data suggest that using the private messaging app Kik is associated with an individual’s photos being used inappropriately. However, we note that the use of Kik should be further investigated in larger studies due to the small number of individuals using this platform in our sample.
From a theoretical perspective, our results suggest the need to better differentiate the relationship between characteristics, and or social media use and the specific type of risk that someone may encounter. From a behaviorist standpoint, extant literature has shown that broad patterns of behavior reflect different fundamental incentives to human action [ 42 , 43 ]. This means that when it comes to online risky situations, including criminal activity, future research should aim to investigate the underlying incentives that guide both youth engagement with certain forms of online behavior and the certain risks that may stem from this. It is critical that forms of online intervention begin to think not just about domain-general aspects of prevention that can be equally beneficial to all (e.g., online safety, online disinhibition, parental rule or regulation), but also domain-specific forms of intervention that may need to be targeted towards certain types of risk that certain sub-groups of young internet users are more likely to be exposed to. These domain-specific forms of intervention must then include not only the specific online domain that individuals may be using (e.g., Kik) and the unique risks that this poses to these groups, but also the deeper psychological phenomena that are driving the online behavior.
The question of coping is important, as well as part of a wider shift from a technologically determinist discourse (of what the internet “does” to children) in favor of recognizing the importance of building children’s digital resilience. Encountering risk could in itself also represent an opportunity–to become more resilient, more digitally literate, and less vulnerable to online risks [ 44 ]. Sustained efforts are required from a variety of stakeholders, including families, schools, and technology industries to effectively safeguard children online in a dynamic technological and social media landscape. Thus, it is important to identify vulnerabilities that expose some children to higher risks so as to focus prevention efforts on those who will most benefit from them.
Future research should continue to investigate how societal factors may explain gender differences and other individual level-characteristics in online exposure to harmful material. We also recommend that future research attempt to understand how parents can reduce children’s vulnerabilities, whose role and supervision was not demonstrated as effective in our study.
Educational initiatives should be tasked with developing programming for youth to address the underlying factors associated with online threats. These initiatives should be designed to develop the critical capability in children to recognize and mitigate the risks encountered by also learning how to self-regulate their behavior and seek help when needed. Children learning how to identify and interpret the impact and potential repercussions of various online behaviors can be influential in the overall safeguarding of youth online; among these potential risks, our study highlighted those posed by chatting with strangers, excessive use of social media platforms, and private messaging apps, and the development of both the technical and emotional competence to self-regulate online behaviors and deal with uncomfortable situations.
Initiatives that build knowledge and awareness about online risks are necessary, as is better awareness of existing disparities in exposure to risks. What this research can lend to the building and framing of educational internet safety initiatives is how to talk about and approach particular risks with students. Each risk online should be approached and discussed uniquely and based on individual risks, avoiding “one-size-fits-all” narratives that assume that all risks and the nature of such risks are the same for every internet user. Internet safety programs, such as those that are most prevalent in the education space like I-SAFE, Common Sense Media, and Netsmartz, could benefit from these findings in their efforts to avoid the use of “one-size-fits-all” approaches. In a multi-year study that evaluated internet safety education programs-such as those above-and their curriculum, researchers identified (among many other findings) that the most effective internet safety education programs need to have different strategies for different internet safety topics [ 45 ]. There is a great need to understand the risk factors and elements that might impact what risks and threats youth may be exposed to on different parts of the internet. Similarly, with many topics within internet safety being complex, the more explicit and clear educational programs can be when talking about risks and where youth might encounter them, the more beneficial it is to the educational goals.
Our findings show that the risks associated with certain online messaging apps are not all-encompassing, but rather narrowly associated with the inappropriate use of ones photos. Furthermore, chatting with strangers on social media has an array of potential associated risks. Understanding this connection and the association of certain platforms, actions, and risks can help inform education programming to give a more direct educational lesson to students to help with improving their self-regulated behaviors, necessary skills, and understanding of the potential risks around them.
Educational organizations and schools should consider our findings when building lessons and their content; learning examples and worksheets can better reflect certain risks and their association with certain behaviors. Additionally, our results should inform the way targeted programming for those who may be disproportionately exposed to risks online should be handled. Our data show that gender is associated with exposure to risky or uncomfortable situations, therefore calling on prevention programs to address these concerns with catered content. Simple changes to existing curricula might include the changing of the subject in particular examples to reflect a more accurate depiction of the research above, or more large-scale changes might be to offer further educational programming for those populations that are shown to be at higher risk of certain exposures.
In considering the findings of this study, several limitations are acknowledged. Since our data were cross-sectional, we lack information on the temporal ordering of our variables, and we cannot assume causality. As a non-random convenience sample, it is important to acknowledge that these results are not necessarily generalizable outside of the sample, and that selection bias may have also influenced our reported results. Furthermore, as some of the exposures and outcomes could be perceived as negative, there is potential for social desirability bias to influence students’ responses. Our study sought to explore associations between risky online events and various behaviors online, demographic characteristics, and parental activities. It was not our intent to create an index to measure an underlying construct of online safety, and as such, no psychometric properties are reported. To our knowledge, there is no extant screening tool for exposure to online risk that has been psychometrically evaluated–future research should aim to address this gap.
Our results provide critical information to practitioners involved in the development of educational initiatives by suggesting the need to identify individual-level characteristics and behaviors that are associated with exposure to online risks in the population targeted by their programs, so to allow them to tailor their initiatives to meet the needs of those more likely to encounter online risks. The results presented in this manuscript also provide a platform for future longitudinal research to further investigate the magnitude of risk associated with these behaviors. Our results should be interpreted under the growing understanding that exposure to risky material online should be understood as a matter of if, not when, and as such, we recommend that future research should investigate how individuals engage with risky online content and the psychological processes which may influence exposure. With the continued increase of youth internet use, the need for reflective, impactful internet safety education is becoming more crucial. Schools are the main means by which a society can address digital safety and citizenship issues for all children within a structured learning environment. To achieve optimal success in this area, it requires teacher training and curriculum materials that are age and gender-appropriate as well as further research on the factors associated with online vulnerability.
The authors of this manuscript would like to acknowledge the work of Marcia A. Testa and Valentina Baccolini on the initial stages of data analysis and manuscript development.
The following are available online at https://www.mdpi.com/article/10.3390/ijerph18115786/s1 . The questionnaire used for data collection is included as a supplementary file. Supplementary Material A: Survey Instrument, Supplementary Material B: Simple Logistic Regression Models.
Conceptualization, E.S. and N.S.; methodology, N.W.H. and M.S.; formal analysis, N.W.H. and M.S.; writing—original draft preparation, E.S., N.W.H., M.S., T.C., and N.S.; writing—review and editing, E.S., N.W.H., T.C., and N.S. All authors have read and agreed to the published version of the manuscript.
Research reported in this publication was supported by the National Institute of Justice under award number 2016-ZA-BX-K001 Evaluation of the Peer to Peer (P2P): Challenging Extremism Initiative and award number 2018–2A-CX-0002 Operation250: An Evaluation of a Primary Prevention Campaign focused on Online Safety and Risk Assessment. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Justice.
The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Harvard T.H. Chan School of Public Health Institutional Review Board (IRB16-1757; 24 January 2017).
Informed consent was obtained from all subjects involved in the study.
Conflicts of interest.
The authors declare no conflict of interests.
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Internet safety research.
When you ask students to think through their own scenarios about online safety, they’re more likely to look out for themselves and others.
The top-line statistics about online safety for kids are troubling: According to a 2019 study by the Center for Cyber Safety and Education, 40 percent of kids in grades 4-8 reported that they connected or chatted online with a stranger. Meanwhile, the Journal of Adolescent Health reports that approximately one in five youth experience unwanted online exposure to sexually explicit material. And the risks just come from strangers: Pew Research Center reports that 59 percent of kids have been bullied or harassed online.
ISTE standards for students don’t quite seem up to the task. While they touch on cyberbullying and data privacy and acknowledge that being a “digital citizen” means learning about the responsibilities and behaviors for interacting in this space, they don’t say a word about access to and sharing of inappropriate and explicit content—including the risk of kids making a naive but devastating misstep themselves.
That’s a big gap. The easy solution would be cut-to-the-chase lectures that treat cyber safety as a standalone topic, but lectures often fail to make meaningful connections with students. In our years of teaching cyber safety to kids, we’ve arrived at what we think is a better approach—one in which cyber safety instruction is woven into subjects such as language arts, social studies, math, PE, and electives like STEAM, health, or theater.
We’ve seen it again and again in our classes: When cyber safety is contextualized and personalized and provides students with an opportunity to create, apply, or synthesize higher-level ideas in their subject-specific areas, the learning is far more profound. Moreover, with a cross-curricular approach, students become more empowered digital citizens who want to share their stories, research the facts, and inform their peers.
Make use of compelling, relevant narratives
Narrative is relevant in most subjects, but we’ve discovered that in the context of ELA, sharing cyber safety narratives with an element of suspense gains students’ attention, gets them thinking critically, and gives them ideas for their own narratives. Start with short, narrative videos such as those from Storybooth , a site that animates real-life stories submitted by 7th-12th graders, including testimonies relevant to cyber safety. BrainPop , which is aimed at elementary teachers and students, also has cyber safety videos.
After watching a video, ask your students to pick a related prompt. For example, if you’ve just watched a video on private and personal information, offer these:
Then they can write a response to the prompt and then turn and talk to a peer or share their responses in a whole group, e.g., “This reminds me of…” or “I wonder what would happen if…”
Or have your students write real or imagined narratives around topics of cyber safety. Start with story prompts such as, “A friend tells you that they are planning to meet up with someone they met through social media and wants you to come along. How do you respond?”
We have students work collaboratively on different story prompts related to cyber safety, then use a jigsaw approach where each group rotates to a completed story prompt with sticky notes to ask questions, share “a-ha” moments, and give peers feedback.
Storytelling with avatars and comics
Comics can both inspire and educate kids . Consider having your students create virtual avatars and comic strips with a tool like Pixton . They can create representations of themselves, classmates, peers, family members, or teachers who will play a role in the narratives they create (with permission, of course, from the represented people—yet another opportunity to model consideration and consent).
For example, kids could envision a scenario in which a high schooler meets someone they think is a peer online, agrees to meet them in person, realizes the gravity of the situation, and flees, and then depict that entire storyline in a comic strip format. The same thing can be done with scenarios related to sharing photos or a bank account number.
With digital comics, students can change backgrounds to build settings, vary facial expressions and body posture of their avatars, and learn technical skills like camera zoom to edit. Digital comics can launch discussions about other important topics, too, like the importance of nonverbal communication and how to support their peers when warning signs of cyber safety arise.
Practice problem-solving with scenarios
When students act out scenarios, they explore possible solutions to complicated problems with their peers. Pose scenarios and ask cooperative groups to act them out. Here's an example:
A group of classmates asks for your help to create a website about a teacher at your school. They want to include inappropriate pictures, images, and comments about her. What do you do?
With practice and trust, your students will likely want to write their own scenarios. Whether you teach health, theater, or creative writing, having kids write a script for a cyber safety situation can bring new energy into your classroom.
Integrate research
When students research facts about cyber safety issues, they build their critical thinking and media literacy skills and gain a clearer understanding of what cyber safety is and what affects it, both in terms of outside influences and their own behavior.
For example, in a social studies class, ask them to consider this scenario: “Your friend tags you in a picture at the beach on Instagram. Other users have begun to make comments that make you feel uncomfortable and don’t seem appropriate to post online. You’re worried that your mom will see it. What should you do?”
As part of the assignment, have students look at legislation related to cyber safety to understand the legal ramifications of bad (and sometimes illegal) behavior; advise them to add links and resources they’ve discovered. Using an online collaboration board such as Padlet, Jamboard, or with Nearpod or Pear Deck, you can also do “think-pair-share” activities rooted in online research.
Relevant research works in math, too, and the objectivity of numbers provides a nice counterbalance to personal narratives. A good infographic, for example, can prompt questions and robust discussion. Because infographics are data-rich, they can reinforce understanding of statistics and probability and help develop quantitative literacy.
Students can also create surveys using Google Forms and collect data from their peers about a cyber safety topic of their choice—they are often more interested in crunching numbers when working with a data set they’ve created or contributed to and care about. Once students have collected and analyzed their data, they can create a Canva infographic of their own, which can also be used for a “notice and wonder” conversation that they lead.
Cyber safety instruction helps students see the connections between good citizenship overall and digital citizenship . Students need to understand the facts, cause and effect, personal responsibility, and the importance of both etiquette and context. Most of all, they need to understand that the lessons of compassion they learn throughout their school years extend into the digital world. One of the best ways to be compassionate towards others (and themselves!) is to be safe online.
An official website of the United States government
Here’s how you know
Official websites use .gov A .gov website belongs to an official government organization in the United States.
Secure .gov websites use HTTPS A lock ( Lock A locked padlock ) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.
https://www.nist.gov/news-events/news/2024/08/us-ai-safety-institute-signs-agreements-regarding-ai-safety-research
These first-of-their-kind agreements between the u.s. government and industry will help advance safe and trustworthy ai innovation for all..
GAITHERSBURG, Md. — Today, the U.S. Artificial Intelligence Safety Institute at the U.S. Department of Commerce’s National Institute of Standards and Technology (NIST) announced agreements that enable formal collaboration on AI safety research, testing and evaluation with both Anthropic and OpenAI.
Each company’s Memorandum of Understanding establishes the framework for the U.S. AI Safety Institute to receive access to major new models from each company prior to and following their public release. The agreements will enable collaborative research on how to evaluate capabilities and safety risks, as well as methods to mitigate those risks.
“Safety is essential to fueling breakthrough technological innovation. With these agreements in place, we look forward to beginning our technical collaborations with Anthropic and OpenAI to advance the science of AI safety,” said Elizabeth Kelly, director of the U.S. AI Safety Institute. “These agreements are just the start, but they are an important milestone as we work to help responsibly steward the future of AI.”
Additionally, the U.S. AI Safety Institute plans to provide feedback to Anthropic and OpenAI on potential safety improvements to their models, in close collaboration with its partners at the U.K. AI Safety Institute.
The U.S. AI Safety Institute builds on NIST’s more than 120-year legacy of advancing measurement science, technology, standards and related tools. Evaluations under these agreements will further NIST’s work on AI by facilitating deep collaboration and exploratory research on advanced AI systems across a range of risk areas.
Evaluations conducted pursuant to these agreements will help advance the safe, secure and trustworthy development and use of AI by building on the Biden-Harris administration’s Executive Order on AI and the voluntary commitments made to the administration by leading AI model developers.
About the U.S. AI Safety Institute
The U.S. AI Safety Institute , located within the Department of Commerce at the National Institute of Standards and Technology (NIST), was established following the Biden-Harris administration’s 2023 Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence to advance the science of AI safety and address the risks posed by advanced AI systems. It is tasked with developing the testing, evaluations and guidelines that will help accelerate safe AI innovation here in the United States and around the world.
Advertisement
Supported by
The maker of ChatGPT is struggling to transform itself into a profit-driven company while satisfying worries about the safety of artificial intelligence.
By Cade Metz and Mike Isaac
Reporting from San Francisco
OpenAI, the often troubled standard-bearer of the tech industry’s push into artificial intelligence, is making substantial changes to its management team, and even how it is organized, as it courts investments from some of the wealthiest companies in the world.
Over the past several months, OpenAI, the maker of the online chatbot ChatGPT, has hired a who’s who of tech executives, disinformation experts and A.I. safety researchers. It has also added seven board members — including a four-star Army general who ran the National Security Agency — while revamping efforts to ensure that its A.I. technologies do not cause serious harm.
OpenAI is also in talks with investors such as Microsoft, Apple, Nvidia and the investment firm Thrive for a deal that would value it at $100 billion . And the company is considering changes to its corporate structure that would make it easier to attract investors.
The San Francisco start-up, after years of public conflict between management and some of its top researchers, is trying to look more like a no-nonsense company ready to lead the tech industry’s march into artificial intelligence. OpenAI is also trying to push last year’s high-profile fight over the management of Sam Altman , its chief executive, into the background.
But interviews with more than 20 current and former OpenAI employees and board members show that the transition has been difficult. Early employees continue to leave, even as new workers and new executives pour in. And rapid growth hasn’t resolved a fundamental question of what OpenAI is supposed to be: Is it a cutting-edge A.I. lab created for the benefit of humanity, or an aspiring industry giant dedicated to profits?
We are having trouble retrieving the article content.
Please enable JavaScript in your browser settings.
Thank you for your patience while we verify access. If you are in Reader mode please exit and log into your Times account, or subscribe for all of The Times.
Thank you for your patience while we verify access.
Already a subscriber? Log in .
Want all of The Times? Subscribe .
If you're on a Galaxy Fold, consider unfolding your phone or viewing it in full screen to best optimize your experience.
Published on Sept. 2, 2024
By: Steve Strauss
While being your own boss has long been a dream for many people, it was not so easy not so long ago. Even just a generation ago, online businesses were not really a thing, and as such, if you wanted to become an entrepreneur, you likely had to start some sort of brick-and-mortar business.
But not today.
The internet changed many things, but one of the best parts of the online digital revolution is that starting a business is easier, faster, and way less expensive than ever before. Gone are the days of renting an expensive office or storefront, hiring staff, paying for overhead, and so on.
Just about anyone can start an affordable business -- often for even under, say, $500. Here are seven examples -- a mix of online and offline businesses. Even offline, physical businesses are cheaper and easier to start nowadays because people can much more easily work out of their homes and marketing can be very inexpensive, or even, as in the case of social media, free.
Why a business credit card could transform your small business.
These business credit cards that offer a convenient and efficient way to separate personal and business expenses, simplifying accounting and tax reporting.
Additionally, business cards can provide valuable perks such as rewards points, cashback, and expense tracking tools, enhancing financial management and the potential to help save money in the long run.
Offer | Our Rating | Welcome Offer | Rewards Program | APR |
---|---|---|---|---|
On Chase's Secure Website. | Our ratings are based on a 5 star scale. 5 stars equals Best. 4 stars equals Excellent. 3 stars equals Good. 2 stars equals Fair. 1 star equals Poor. We want your money to work harder for you. Which is why our ratings are biased toward offers that deliver versatility while cutting out-of-pocket costs. | Earn $750 bonus cash back Earn $750 bonus cash back after you spend $6,000 on purchases in the first 3 months from account opening. | Earn unlimited 1.5% cash back on every purchase Earn unlimited 1.5% cash back on every purchase made for your business | 0% Intro APR on Purchases Purchases: 0% Intro APR on Purchases, 12 months Balance Transfers: N/A 18.49% - 24.49% Variable |
On Chase's Secure Website. | Our ratings are based on a 5 star scale. 5 stars equals Best. 4 stars equals Excellent. 3 stars equals Good. 2 stars equals Fair. 1 star equals Poor. We want your money to work harder for you. Which is why our ratings are biased toward offers that deliver versatility while cutting out-of-pocket costs. | Earn up to $750 bonus cash back Earn $350 when you spend $3,000 on purchases in the first three months and an additional $400 when you spend $6,000 on purchases in the first six months after account opening. | Earn 5% cash back in select business categories Earn 5% cash back on the first $25,000 spent in combined purchases at office supply stores and on internet, cable and phone services each account anniversary year. Earn 2% cash back on the first $25,000 spent in combined purchases at gas stations and restaurants each account anniversary year. Earn 1% cash back on all other card purchases with no limit to the amount you can earn. | 0% Intro APR on Purchases Purchases: 0% Intro APR on Purchases, 12 months Balance Transfers: N/A 18.49% - 24.49% Variable |
The demand for VAs is growing, particularly among businesses that need help with day-to-day tasks but are not ready to hire full-time staff. Starting this business requires little more than a computer, internet access, and some apps and software.
Along the same lines, while we all know that having a social media presence is vital in today's always-on, too-connected world, many people, small businesses especially, can't keep up with it. Either they don't have the content creation chops, or the technical know-how, time, or interest to manage their social media.
That's where you come in.
You can offer services like content creation, post scheduling, and performance analysis. The email marketing company Vertical Response found in a survey that 43% of small businesses spend about six hours a week on social media management. Get yourself a few clients, and you will have a full schedule (and a fat bank account!)
Coaches are all the rage these days. Whether it is selling a course or one-on-one coaching packages, online coaches sell and teach everything from yoga and weight loss to money management and online marketing. Simply set up a website, market yourself with some pay-per-click ads, and away you go.
Similarly, tutoring and teaching online has also exploded in recent years. My own daughter spent the year after college teaching Chinese kids how to speak English via the internet. She made bank (although sleep was not part of that bank, as she had to wake up at 3:00 a.m. to fit their schedule). Online tutoring platforms can be found with a simple Google search.
Now let's pivot and look at a couple of offline businesses you could also start on the cheap.
We all love our pets, that's for sure (although I am sorry to report that our goldendoodle Hazel is, statistically speaking, the very best dog in the world). And as pet ownership continues to grow, so does the need for reliable pet care services.
If you love animals, pet sitting can be a rewarding and flexible business. Apps like Rover allow you to advertise your business and it handles the administrative side of things. Your responsibilities may include feeding, walking, grooming, and playing with pets while their owners are out of town or at work. The start-up costs are low, typically limited to basic supplies and insurance.
Starting a house or commercial office cleaning business can be a lucrative and affordable option. This business requires minimal initial investment and skill, just some cleaning supplies and basic equipment. Undercut the competition with your prices and you are in business (literally).
Finally, if you are good at fixing things, this old standby remains a very viable option. From home repairs to painting, plumbing, and electrical work, there is no shortage of things around the house for a handyperson to fix. Start-up costs are minimal, and advertising on sites like Craigslist is very cheap, often free.
So there you have it. Entrepreneurship on a dime is within your grasp. Whether it is managing social media or providing cleaning services, starting small and cheap really is quite possible.
Steve Strauss is the president of a boutique content company, The Strauss Group, and is a bestselling small business author and columnist. He can be reached at www.MrAllBiz.com, or at [email protected] .
Share this page
We're firm believers in the Golden Rule, which is why editorial opinions are ours alone and have not been previously reviewed, approved, or endorsed by included advertisers. The Ascent, a Motley Fool service, does not cover all offers on the market. The Ascent has a dedicated team of editors and analysts focused on personal finance, and they follow the same set of publishing standards and editorial integrity while maintaining professional separation from the analysts and editors on other Motley Fool brands.
Related Articles
By: Cole Tretheway | Published on June 7, 2024
By: Lyle Daly | Published on June 5, 2024
By: Christy Bieber | Published on June 5, 2024
By: Lyle Daly | Published on June 4, 2024
The Ascent is a Motley Fool service that rates and reviews essential products for your everyday money matters.
Copyright © 2018 - 2024 The Ascent. All rights reserved.
IMAGES
VIDEO
COMMENTS
1. Introduction. Children today spend substantial time online for education, entertainment, communication, and social interaction. Being online offers a wealth of new opportunities for realising children's rights, but also poses new risks for their violation and abuse (Committee on the Rights of the Child, 2021).To be safe online, children need education about online opportunities and risks ...
long-term impacts of existing cyber-safety program s and approaches. Purpose. The purpose of this literature review is to better understand h ow children and young people engage with the. digital ...
Youth Internet Safety: Risks, Responses, and Research Recommendations. As Internet use by children and teenagers increases, so do concerns about their online safety. Providing a safe environment ...
June 30, 2024. We research online 'misogynist radicalisation'. Here's what parents of boys should know. Steven Roberts, Monash University and Stephanie Wescott, Monash University. Parents ...
This study first analyzes online safety research trends, followed by categorizing the types of online risks within the "4Cs framework". Moreover, it conducts a case study on Australia, Canada, and the United Kingdom (UK), the three most advanced countries in terms of online safety-related policies for children and youth . It attempts to ...
As part of her research, she holds workshops with teens to involve them in co-designing online safety interventions. Though this program, called Teenovate, is in the early stages, some ideas have ...
The concept of online safety is sometimes divided into the 3Cs: concerns regarding content (illegal or inappropriate images, text or video), contact (by unknown or unsafe parties, particularly regarding children) and conduct (such as grooming, bullying or revenge porn) [].This schema focuses on threats by 'outsiders' to vulnerable users such as children who are framed as potential victims.
Online safety awareness remains a pressing issue as the number of cyberattacks has been increasing in number and in the complexity of the attacks [6]. While the low level of online safety awareness is well-documented, there is a lack of comprehensive discussion on the complex intra- and interpersonal human factors that play into this phenomenon.
Using terms "internet safety education" and "digital citizenship," a Google search identified 12 multi-topic safety programs. Review articles were identified via Google Scholar for six forms of online harm to youth that have been targeted by many of these programs: cyberbullying (19 articles); online sexual exploitation (23 articles); sexting (19 articles); online fraud, hacking, and ...
Research Article. Digital crime, trauma, and abuse: Internet safety and cyber risks for adolescents and emerging adults in the 21 st century. Yok-Fong Paat a Department of Social Work, The University of Texas at El Paso, El Paso, Texas, USA Correspondence [email protected]
Internet use is nearly ubiquitous among US youth; risks to internet use include cyberbullying, privacy violations and unwanted solicitation. Internet safety education may prevent these negative consequences; however, it is unclear at what age this education should begin and what group is responsible for teaching this topic. Surveys were distributed to key stakeholders in youth safety education ...
1 Youth Internet Safety: Risks, Responses, and Research Recommendations Adina Farrukh, Rebecca Sadwick, and John Villasenor I. IntroductIon A s Internet use by children and teenagers increases, so ...
1.1. Adolescents' Exposure to Risky Online Situations. Risky online situations that adolescents may encounter have been classified into three groups: (1) "content risk," referring to user exposure to possible harmful contents, (2) "contact risk," referring to user activities and communications with known or unknown individuals that carry a potential threat and (3) "commercial risk ...
In 2017, Pinter, Wisniewski, Xu, Rosson, and Caroll (2017) conducted a literature review on adolescent online safety, reporting important trends by synthesizing 132 peer-reviewed publications. Among other thought-provoking findings, the researchers report that 66% of the reviewed articles were focused on cyberbullying.
Rule 2: Digital security is rarely the weakest link. The second rule is to remember that most attacks take place offline and that the most effective ones are often the simplest. At an airport several years ago, 1 author heard a professor of computer science try to reset an online account over the phone.
However, the impact of children's subjective versus objective knowledge on their perceived online safety and attitudes towards e-safety education remain unclear. Questionnaires were used to assess children's (N = 329, aged 8-11 years) perceived online safety, subjective and objective knowledge of online safety/dangers, and attitudes to e ...
referred to as " online safety ," " digital safety," o r "e-saf ety," this concept i s associated. both with the risks in dividuals face onl ine and with the wa ys they can protect th ...
Study of the Internet and Youth "At Risk Behaviors" (Rochester Institute of Technology, September, 2008) Cox Communications Teen Internet Safety Survey, Wave II--in Partnership with the National Center for Missing and Exploited Children and John Walsh (2007) The Enough Is Enough mission is to make the Internet Safer for Children and Families.
Over 30 years ago, Mason (Citation 1986) voiced ethical concerns over the protection of informational privacy, or "the ability of the individual to personally control information about one's self" (Stone et al., Citation 1983), calling it one of the four ethical issues of the information age.Since the 1980s, scholars have remained concerned about informational privacy, especially given ...
The top-line statistics about online safety for kids are troubling: According to a 2019 study by the Center for Cyber Safety and Education, 40 percent of kids in grades 4-8 reported that they connected or chatted online with a stranger. Meanwhile, the Journal of Adolescent Health reports that approximately one in five youth experience unwanted ...
Surveys in 2021 and 2022 showed that 22% and 38% of respondent researchers, respectively, received threats of violence after talking to the media about COVID-19. Similarly, a worldwide survey by Global Witness published in 2023 found that 39% of respondents experienced online harassment or abuse related to their climate research. And as documented in Scholars at Risk's annual monitor ...
Online Safety; Online Safety. en español: Seguridad en línea. Medically reviewed by: Elana Pearl Ben-Joseph, MD. Listen. ... take pictures, do our homework, do research, find out the latest news, and shop. But besides the millions of sites to visit and things to do, going online offers lots of ways to waste time — and even get into trouble.
Each company's Memorandum of Understanding establishes the framework for the U.S. AI Safety Institute to receive access to major new models from each company prior to and following their public release. The agreements will enable collaborative research on how to evaluate capabilities and safety risks, as well as methods to mitigate those risks.
By investigating the UK's Online Safety Bill, comparing it with similar interventions, and considering the political impact of different digital tools for moderation, this perspective aims to inform the current policy debate by combining technical and political insight. ... Self-harm and suicide: the White Paper cites academic research that ...
As feminist academics invested in producing usable research, GMMP can - and does - help lobby for change in our work with media organisations, offering an important opportunity for longitudinal and international comparisons due to the relative stability of its methodology across time and place.
1 INTRODUCTION. As the effects of climate change continue to be felt, severe wildfires 1 have increased in occurrence, duration, and intensity. 1 Recent mega-fires across the globe provide evidence of the scale and impacts of this hazard 2-4 which continue to be one of the costliest forms of natural disasters globally from a financial and a human perspective. 5-13 This trend has continued ...
Last month, Dr. Schulman, the co-founder who helped oversee OpenAI's new safety efforts, also resigned from the company, saying he wanted to return to "hands-on" technical work. He also ...
Conducting research; Content creation; And much more; The demand for VAs is growing, particularly among businesses that need help with day-to-day tasks but are not ready to hire full-time staff ...
A Safety and Health Research Forum. A Joint Publication of the National Safety Council and Elsevier. The Journal of Safety Research is a multidisciplinary publication that provides for the exchange of scientific evidence in all areas of safety and health, including traffic, workplace, home, and …. View full aims & scope. National Safety Council.
The California Health Sciences University College of Osteopathic Medicine building in Clovis' Research and Technology Park is at center in the drone image made on Wednesday, Sept. 27, 2023 ...