Why do things get misused? Why are things abused? Are the things the problem? Is there anything that cannot be misused or abused? What makes the human mind [check] misuse and abuse? How has human society mitigated both across objects and events? What are priority areas against misuses and abuses? Where does AI place?
If an individual misuses an elevator—going up and down so frequently, within a floor, what is the cost of that misuse? If a pen is used to open a package rather than to write, how much of a misuse is that? When does misuse start to become a problem?
Misuses can be graded by scale, and sometimes by intensity or by effort to correct the error. Though misuses and abuses for physical objects still pose risks in different ways to society, one of the major risk areas for misuses and abuses is digital.
The possibility for scale, the [relatively] small time required, and some of the broader damages possibly make digital a platform that can be abused and misused differently from the physical [so to speak].
Though it is easy to learn to use a phone or social media, there were outcomes within digital that required skills that AI can now cover. This means that digital—with its risks, then AI, with its capabilities—presents a field for misuses and abuses, globally.
There are already stories of fake images of students, fake audios of people, fake texts, and fake videos. There are victims in all these cases, and there might be several unknown and unannounced cases. There could also be an impact on the long term in several cases.
AI is supposed to be a technology for good, but with the possibility for misuses and abuses, there would be several angles of risks it would carry.
How can these risks be prevented? The problem of AI safety is a problem for any source that people could be affected by. This means that school districts where AI has caused problems for students would need their own AI safety institutes. States would also do, counties and then university systems, to have their own fronts against the risks and threats of AI.
Even if it is not a complex institute, they need a team to keep asking the question, how can misuses of AI be combated in ways to protect their organization and by large the sector?
There are lots of conceptual possibilities outside of some of the general approaches in AI alignment that could be sought. This can become a major angle towards progress, against delayed reaction, while AI becomes a tool that can be used for serious damage against many young people.
There is a recent story on The Korea Herald, S. Korea’s AI safety institute officially launches, stating that, “South Korea’s artificial intelligence safety institute officially opened Wednesday to address various risks associated with the rapidly evolving technology, the science ministry said. The new AI Safety Institute follows the AI Seoul Summit held in May, where leaders from South Korea, Britain, and eight other nations adopted a joint declaration on promoting safe, innovative, and inclusive AI, according to the Ministry of Science and ICT. Located in Pangyo, just south of Seoul, the institute will lead the country’s research on risks involving AI, such as misuse and loss of control, while serving as a hub for industry, academia, and research institutes studying AI safety. It will also participate in the international network of AI safety institutes to foster global discussions on the agenda, the ministry said.”
There is a recent report by Times Higher Education, ChatGPT: student AI cheating cases soar at UK universities, stating that, “New figures obtained by Times Higher Education indicate that suspected cases of students illicitly using ChatGPT and other AI-assisted technologies in assessments skyrocketed in the past academic year while the number of penalties – from written warnings and grade reductions to the refusal of credits and failure of entire modules – has also increased dramatically.
At the University of Sheffield, there were 92 cases of suspected AI-related misconduct in 2023-24, for which 79 students were issued penalties, compared with just six suspected cases and six penalties in 2022-23, the year in which ChatGPT was launched. At Queen Mary University of London, there were 89 suspected cases of AI cheating in 2023-24 – all of which led to penalties – compared with 10 suspected cases and nine penalties in the prior 12 months. At the University of Glasgow there were 130 suspected cases of AI cheating in 2023-24 – with 78 penalties imposed so far and further investigations pending, compared with 36 suspected cases and 26 penalties in 2022-23.”
—
iStock image
