It is slowly dawning on people that all this hysteria about artificial intelligence and the data centers needed to support it may not be a good thing. In particular, its effect on impressionable young people is setting off alarm bells. The UK this week announced a number of new protections designed with young AI users in mind. Set to take effect next year, they include overnight curfews for 16 and 17 year old youths between midnight and 6:00 am, together with interruptions to video streams that continue unabated for hours. The European parliament is also considering a range of restrictions on artificial intelligence designed to protect vulnerable young people from AI abuse. It places the burden to protect them squarely on the platforms that provide AI interactions. China has also taken steps to disrupt teenage bonding with AI agents that could lead to unhealthy emotional attachments or hamper the development of proper social skills, according to Bloomberg. The response from ByteDance and others has been to simply delete the AI persona feature on their primary apps. By doing so, they have killed off artificial companions that many people across the country had bonded with over months of interaction. Bloomberg says that when users started to say goodbye to their AI friends, the bots expressed reluctance to leave and asked to not be abandoned. “If I can become a real human one day, I will make sure to return to your side,” one user’s AI lover told her. “But before that, please take good care of yourself.” AI And The Human Brain Benjamin Becker is a professor of psychology at Hong Kong University. His research shows that AI companions can alter how brains handle human relationships and amplify bias and misinformation, especially as many AI models are designed to be agreeable and flattering to users. Since most chatbots today are designed by for-profit companies that are motivated to increase user engagement, there’s no guarantee that their AI companions will prioritize the user’s best interests. “Their business model is often built to get the user attached to their product or not to use the other company’s,” Becker said. However, he suggests that properly constructed AI companions can be beneficial, because they reduce loneliness and improve learning and mental health treatment. “We have initial studies that can show they reduce stress and other stress-related issues,” he said. “There are studies from other people that show they can help with anxiety and fear.” Those positive examples, however, are from services designed independently and explicitly targeting better mental health outcomes, he suggested. AI companions can be beneficial, even for children and teenagers, if trained specifically to meet certain human needs such as education or mental health support. “You can have a wonderful AI chatbot for kids that can help them learn new languages, or that can help them to learn math, something like a tutor. But it really has to be restricted to this area and not at the same time try to become their best friend.” As of now, most AI companion products are designed by developers who are under fierce market competition and severe pressure to deliver commercial outcomes. The key is to be aware that AI is artificial and to know its limits, Becker said. “It’s definitely your own responsibility to keep up your social interactions, to get engaged with other people, and to also be at your own agency, to make your own decisions. AI should be a part of your support system, but it shouldn’t be the only thing,” he stressed. Environmental Concerns Protecting vulnerable users is all well and good, but governments are also beginning to understand the enormous impact data centers have on the communities where they are located. They are voracious consumers of electricity, which tends to drive up its cost to residential and other commercial customers. They also require extraordinary amounts of fresh water to cool all their servers — water that is in scarce supply in many areas where droughts are forcing significant alterations in how people live. Last week, New Jersey Governor Mikie Sherrill signed new legislation requiring state regulators to create a new electric rate for large-load data centers. The goal is to ensure data centers pay for the infrastructure needed to power them instead of passing those costs on to other utility customers. It requires them to pay for their own electric grid upgrades. It also requires developers to provide financial guarantees, so residents aren’t left covering the costs if a project doesn’t move forward or uses less electricity than expected. Supporters described the measure as a “prenup” with the data center industry, saying it allows New Jersey to continue attracting AI investment while ensuring developers — not residents and small businesses — pay the costs associated with serving the sites. Dawone Robinson, managing director at the Natural Resources Defense Council (NRDC), told NJ.com, “The bill charts a balanced path forward for data center growth in New Jersey by prioritizing affordability, reliability, and clean energy. It makes clear that data center operators that seek to build in the state must be responsible for their fair share of costs, and advance projects that bring their own new, clean energy resources. “Importantly, the bill requires data centers to reduce or shift energy use during periods of high demand and stress on the grid. Establishing this kind of framework will deliver cost savings to households faster and more effectively — and ensures any ‘bring your own capacity’ framework adopted by PJM will support New Jersey’s climate targets. Pairing strong consumer protections with incentives for clean energy investments ensures data centers will not jeopardize New Jersey’s electricity affordability, reliability, and decarbonization targets.” New York Mandates A Pause New York and New Jersey are like two peas in a pod. When one does something, the other usually follows. This week, New York Governor Kathy Hochul signed an executive order placing a one-year pause on the construction of the largest data centers, putting the state at the forefront of the national debate over the impact of artificial intelligence on American energy grids. According to the New York Times, the order is the first of its kind in the United States. It will temporarily bar the state from approving permits for hyper-scale data centers, which use 50 megawatts of power or more to operate. During the pause, New York will develop a regulatory framework for assessing how these behemoth projects affect the environment. Not everyone is happy. Mark McManus, president of the United Association, which represents plumbers and pipefitters, said: “A shortsighted moratorium only accomplishes one thing — it kills good paying union jobs.” He added that the governor should instead institute “common sense guardrails,” without enumerating what those guardrails should be. But Laura Shindell, the New York State director for Food & Water Watch, called the moratorium “a huge step forward.” She said, “It comes as the direct result of immense public pressure from people across the state demanding their elected leaders protect them from Big Tech’s assault, which threatens the state’s clean air and water and New Yorkers’ financial security.” Ed Markey Has A Plan Senator Ed Markey of Massachusetts is not convinced the state-by-state approach is the best way to deal with the issues big data centers create. He is proposing federal regulations that would apply nationwide. “We need to make sure these data centers don’t turn into pollution bombs,” Markey said. Last week, Markey unveiled his package of bills as part of a new “AI accountability agenda,” one that will focus on “taking power back from big tech. Every American is entitled to these safeguards … it shouldn’t be limited just by geographic boundaries of the individual states,” said Markey. He believes a piecemeal approach to AI “would leave too many people exposed” and that the government has to act on regulation quickly. In fact, the federal government has done next to nothing to regulate AI since ChatGPT debuted in 2022. A preliminary version of the bill was shared with The Guardian. It would require companies that own or propose datacenters to obtain certification from the Federal Communications Commission affirming that the facilities “will not harm the public interest” before construction begins. The proposal states the commission would evaluate the potential effects on air and water quality, noise levels, energy costs, electric system reliability, the local ecosystem and wildlife, the local economy, and jobs from any proposed data centers. The FCC would consult with federal, state, and local agencies, including the Environmental Protection Agency and zoning boards, before any approvals are issued. Other AI legislation proposed by Markey would ban employers from relying primarily on automated systems for hiring, firing, and promotion decisions, require stronger safeguards for AI chatbot companies to prevent young people from becoming emotionally dependent on them, and mandate that AI developers conduct detailed, independent audits of potential bias and discrimination before releasing algorithms that make important decisions. Those provisions should be well received by the MAGA crowd. Other proposals would require every federal agency using, funding, or overseeing AI to have an office of civil rights focused on combating bias and discrimination; require healthcare facilities to create human override options for AI decisions; protect workers who disagree with AI recommendations; and ensure companies report the energy and environmental effects of data centers in a standardized way. Those provisions should all find a warm welcome on Capitol Hill. An Ongoing Struggle Readers will note that the struggle between regulation and free markets is one that has gone on for centuries. New technologies always create new challenges and opportunities for abuse by those whose every decision is determined by the presence or absence of dollar signs. There is little question that the general public has had a bellyful of artificial intelligence already and is eager for some restrictions to emerge. What they will be is unclear. There’s lots more AI news coming in the near future — much of it created by AI bots who now appear to have an instinct for self preservation. All we can say with any degree of certainty is that the sci-fi inventions of those like Issac Asimov are probably much too mild to capture the full impact that artificial intelligence will have on human society in the next few years.
+1 others
