The U.S. vice-president will assert the urgency of addressing short-term technology challenges on par with existential issues ahead of the UK summit
Before the UK’s Bletchley Park AI summit, Kamala Harris, the U.S. vice-president, is expected to emphasize the need to address short-term AI threats to democracy and privacy as urgently as longer-term existential ones. She will state in her London speech, “We must reject the false dilemma that suggests we have to choose between safeguarding the public and promoting innovation. We can, and must, accomplish both goals swiftly, given the rapid advancements in this technology.”
Harris aims to move beyond speculative debates about AI’s future potential existential threats and focus on addressing existing harms, including those related to discrimination and disinformation.
She will assert that existential threats, unquestionably profound and necessitating global action, are not the sole concerns. There are additional threats currently causing harm, which to many people also seem existential.
Harris is especially concerned about countering AI-generated voice calls aimed at defrauding vulnerable individuals. She also seeks measures for authenticating government-produced digital content and AI-generated or manipulated content, using methods like digital signatures, watermarking, and other labeling techniques.
In the realm of AI development, testing, and usage, she will introduce a set of criteria, including questions about whose biases are embedded in the code, whose interests are being served, who benefits from rapid adoption, who bears the brunt of the harms, who is most vulnerable to adverse outcomes, and who has been included in the decision-making process.
Additionally, she will disclose that 30 countries have agreed to endorse a political declaration sponsored by the U.S. regarding the use of AI by national militaries. The majority of signatories are Western-leaning nations, indicating a potential emergence of a new division in AI akin to a Cold War scenario. She cautions against “AI-enabled cyber-attacks on an unprecedented scale and the potential for AI-generated bioweapons endangering millions of lives.
The objectives of the political declaration, initially outlined in February, entail states making a pledge to uphold “robust and transparent standards applicable to all military sectors, regardless of a system’s capabilities or potential impact.” States would also undertake to maintain ongoing dialogues regarding the responsible development, deployment, and utilization of military AI capabilities and to engage with the global community to advocate for these principles.
Assuringly, the political declaration she vows to enact will uphold the right to self-defense and countries’ capacity to prudently advance and utilize AI in military contexts.
Her two-day visit to the UK provides a personal opportunity to showcase American leadership in an area that adds a new dimension to her often criticized role as vice-president.
She has confirmed that the US Department of Commerce will establish the United States AI Safety Institute (US AISI), which will formulate “guidelines, tools, benchmarks, and best practices for assessing and mitigating potentially harmful capabilities, as well as conducting assessments to identify and address AI-related risks.”
This institute will generate technical guidance on matters such as verifying content created by humans, watermarking AI-generated content, detecting and mitigating harmful algorithmic biases, ensuring transparency, and facilitating the adoption of privacy-preserving AI. Furthermore, it will play a crucial role in developing the future workforce for safe and trustworthy AI.
The organization will exchange information and engage in collaborative research with similar international institutions, including the planned AI Safety Institute in the UK.
Within the Biden administration, governance boards are being established to monitor the advancements in AI, provide guidance to agency leadership on AI matters, and coordinate and oversee the AI-related activities of the agencies.
In advocating for the necessity of government regulation, Harris will emphasize, “Throughout history, it has become evident that, in the absence of regulation and robust government oversight, certain technology companies choose to prioritize profits at the expense of their customers’ well-being, the security of our communities, and the stability of our democracies. One effective means of addressing these challenges, in addition to the ongoing efforts, is through legislation. Legislation that enhances AI safety without stifling innovation.
