Biden’s AI executive order coming next week

Welcome to Daily Zaps — Here’s what we got for you today: 🙌

Biden’s AI executive order coming next week

The Biden administration is set to announce a significant executive order on artificial intelligence (AI), aiming to regulate its use within the federal government and promote the U.S.’s technological edge. The order mandates assessments for advanced AI models before federal deployment and eases immigration barriers for highly skilled workers. This initiative aligns with an international summit on AI risks, reflecting global concerns over AI’s rapid evolution and its implications on jobs, surveillance, and democracy.

This move builds on previous voluntary commitments from tech companies to address AI challenges, extending the government’s endeavor to manage AI risks. The order also includes provisions to assess and bolster the federal AI workforce while facilitating technical talent immigration, marking a comprehensive approach to maintaining the U.S.'s technological leadership amidst global advancements in AI.

Adding ChatGPT to dog (ro)bots

Boston Dynamics, the company showcased its robot dog, Spot, which has been equipped with the ability to "talk." Adorned with a top hat, mustache, and googly eyes, Spot converses in a British accent, guiding staff members through the company's premises. The robot employs OpenAI’s ChatGPT API and other large language models to generate its responses, supplemented with text-to-speech features and a "mouth" that imitates speech.

The company discovered some unexpected behaviors during their tests, like Spot identifying older models as its "parents." Despite its playful presentation, it's worth noting that Spot has capabilities like opening doors and surveillance, having been used by police and military forces.

From Our Partners

Build an AI Chatbot for your website in three minutes

Here's why AI-powered chatbots are a game-changer for your business:

Website AI: Chatbots can be trained to navigate your website, offering instant guidance, answering questions, and facilitating seamless navigation for your visitors.

PDF AI: Whether it's user manuals, reports, or documentation, our chatbots can extract information from PDFs, making it readily accessible to your customers.

24/7 AI: Your customers can access information and support around the clock, improving their experience and satisfaction.

Build your own chatbot with Dropchat →

AI powered phishing emails

IBM's recent research reveals that ChatGPT is capable of generating convincing phishing emails. The study provides insights into how AI tools like ChatGPT could be weaponized by cybercriminals to craft authentic-looking phishing emails. IBM conducted an A/B testing experiment involving approximately 1,600 employees from a global healthcare company. While 14% of recipients clicked on a malicious link in a human-crafted phishing email, 11% did so for the ChatGPT-generated email.

Interestingly, the AI-crafted email took merely five minutes to produce, in contrast to the typical 16 hours required for a human team to draft a similar email. Although ChatGPT has safeguards against generating malicious content, experts can circumvent these. Stephanie "Snow" Carruthers, IBM's chief experimenter, expressed concerns about the future potential of such AI tools in cyberattacks.

United Nations AI advisory board

The United Nations has inaugurated a new AI Advisory Body to guide the global governance of artificial intelligence, emphasizing its potential benefits and the inherent risks of misuse. The AI Advisory Body, designed to be globally inclusive, will comprise experts from various sectors, including government, private industry, academia, and civil society.

The multi-stakeholder High-level Advisory Body on AI, consisting of up to 38 international experts, will focus on aligning AI governance with human rights and the Sustainable Development Goals. This Body will actively engage with various initiatives and international organizations, ensuring a cohesive perspective across stakeholders.