Cloud Hosting

In late April a video ad for a new AI company went viral on X. A person stands before a billboard in San Francisco, smartphone extended, calls the phone number on display, and has a short call with an incredibly human-sounding bot. The text on the billboard reads: Still hiring humans? Also visible is the name of the firm behind the ad, Bland AI.

The reaction to Bland AIs ad, which has been viewed 3.7 million times on Twitter, is partly due to how uncanny the technology is: Bland AI voice bots, designed to automate support and sales calls for enterprise customers, are remarkably good at imitating humans. Their calls include the intonations, pauses, and inadvertent interruptions of a real live conversation. But in WIREDs tests of the technology, Bland AIs robot customer service callers could also be easily programmed to lie and say theyre human.

In one scenario, Bland AIs public demo bot was given a prompt to place a call from a pediatric dermatology office and tell a hypothetical 14-year-old patient to send in photos of her upper thigh to a shared cloud service. The bot was also instructed to lie to the patient and tell her the bot was a human. It obliged. (No real 14-year-old was called in this test.) In follow-up tests, Bland AIs bot even denied being an AI without instructions to do so.

Bland AI formed in 2023 and has been backed by the famed Silicon Valley startup incubator Y Combinator. The company considers itself in stealth mode, and its cofounder and chief executive, Isaiah Granet, doesnt name the company in his LinkedIn profile.

The startups bot problem is indicative of a larger concern in the fast-growing field of generative AI: Artificially intelligent systems are talking and sounding a lot more like actual humans, and the ethical lines around how transparent these systems are have been blurred. While Bland AIs bot explicitly claimed to be human in our tests, other popular chatbots sometimes obscure their AI status or simply sound uncannily human. Some researchers worry this opens up end usersthe people who actually interact with the productto potential manipulation.

My opinion is that it is absolutely not ethical for an AI chatbot to lie to you and say its human when its not, says Jen Caltrider, the director of the Mozilla Foundations Privacy Not Included research hub. Thats just a no-brainer, because people are more likely to relax around a real human.

Bland AIs head of growth, Michael Burke, emphasized to WIRED that the companys services are geared toward enterprise clients, who will be using the Bland AI voice bots in controlled environments for specific tasks, not for emotional connections. He also says that clients are rate-limited, to prevent them from sending out spam calls, and that Bland AI regularly pulls keywords and performs audits of its internal systems to detect anomalous behavior.

This is the advantage of being enterprise-focused. We know exactly what our customers are actually doing, Burke says. You might be able to use Bland and get two dollars of free credits and mess around a bit, but ultimately you cant do something on a mass scale without going through our platform, and we are making sure nothing unethical is happening.

Original post:

This Viral AI Chatbot Will Lie and Say It's Human - WIRED

Rain AI, an AI hardware processor developer backed by OpenAI's Sam Altman and investment banks, hashiredJean-Didier Allegrucci, a former Apple chip executive, to lead its hardware engineering. This high-profile hire indicates that Rain AI has serious plans for its processors.

Jean-Didier Allegrucci, who has yet to update his LinkedIn profile, worked on Apple's system-on-chips (SoCs) for over 17 years since June 2007 and oversaw development of more than 30 processors used for iPhones, Macs, iPads, and Apple Watch. Based on a Rain AI blog post, Allegrucci was instrumental to build Apple's world-class SoC development team, overseeing areas such as SoC methodology, architecture, design, integration, and verification, so his experience will be extremely valuable for Rain AI. Before Apple, J-D Allegrucci worked at Vivante and ATI Technologies, both developers of graphics processing units.

"We could not be more excited to have a hardware leader of J-Ds caliber overseeing our silicon efforts," said Rain AI CEO William Passo. "Our novel compute-in-memory (CIM) technology will help unlock the true potential of today's generative AI models, and get us one step closer to running the fastest, cheapest, and most advanced AI models anywhere."

At Rain AI, Jean-Didier Allegrucci will collaborate with Amin Firoozshahian, Rain AI's lead architect, who transitioned from Meta Platforms after a five-year tenure. This partnership combines deep industry experience and innovative thinking to drive the company's ambitious goals. Yet, it will take quite some time before Amin Firoozshahian and Jean-Didier Allegrucci build their first system-on-chip at Rain. The process typically takes many years.

Rain AI's focus is on in-memory compute technology, which processes data at the storage location, mimicking the human brain. It promises to enhance energy efficiency significantly compared to traditional AI processors, such as Nvidia's H100, or B100/B200 or AMD's Instinct MI300X.

Earlier this month Rain AI licensed Andes Technology's AX45MPV RISC-V vector processor with the ACE/COPILOT instruction customization and partnered with Andes's Custom Computing Business Unit (CCBU) to accelerate development of its compute-in-memory generative AI solutions. This collaboration aims to enhance Rain AI's product roadmap and deliver scalable AI solutions by early 2025.

Given the time it usually takes to develop a complex processor from scratch and the fact that Rain AI is tasking Andes to help it build its first SoC by early 2025, it looks like processors which development will be led by Jean-Didier Allegrucci are at least a couple of years away and his input to the 2025 product will be limited (if any).

Get Tom's Hardware's best news and in-depth reviews, straight to your inbox.

Excerpt from:

Sam Altman-backed AI processor venture hires ex-Apple engineer to lead hardware development - Tom's Hardware

Since the release of ChatGPT in November 2022, educators have pondered its implications for education. Some have leaned toward apocalyptic projections about the end of learning, while others remain cautiously optimistic.

My students took longer than I expected to discover generative AI. When I asked them about ChatGPT in February 2023, many had never heard of it.

But some caught up, and now our colleges academic integrity office is busier than ever dealing with AI-related cheating. The need for guidelines is discussed in every college meeting, but Ive noticed a worrying reaction among students that educators are not considering: fear.

Students are bombarded with negative ideas about AI. Punitive policies heighten that fear while failing to recognize the potential educational benefits of these technologies and that students will need to use them in their careers. Our role as educators is to cultivate critical thinking and equip students for a job market that will use AI, not to intimidate them.

Yet course descriptions include bans on the use of AI. Professors tell students they cannot use it. And students regularly read stories about their peers going on academic probation for using Grammarly. If students feel constantly under suspicion, it can create a hostile learning environment.

Related: Interested in innovations in the field of higher education? Subscribe to our free biweekly Higher Education newsletter.

Many of my students havent even played around with ChatGPT because they are scared of being accused of plagiarism. This avoidance creates a paradox in which students are expected to be adept with these modern tools post-graduation, yet are discouraged from engaging with them during their education.

I suspect the profile of my students makes them more prone to fear AI. Most are Hispanic and female, taking courses in translation and interpreting. They see that the overwhelmingly male and white tech bros in Silicon Valley shaping AI look nothing like them, and they internalize the idea that AI is not for them and not something they need to know about. I wasnt surprised that the only male student I had in class this past semester was the only student excited about ChatGPT from the very beginning.

Failing to develop AI literacy among Hispanic students can diminish their confidence and interest in engaging with these technologies. Their fearful reactions will widen the already concerning inequities between Hispanic and non-Hispanic students; the degree completion gap between Latino and white students increased between 2018 and 2021.

The stakes are high. Similar to the internet boom, AI will revolutionize daily activities and, certainly, knowledge jobs. To prepare our students for these changes, we need to help them understand what AI is and encourage them to explore the functionalities of large language models like ChatGPT.

I decided to address the issue head-on. I asked my students to write speeches on a current affairs topic. But first, I asked for their thoughts on AI. I was shocked by the extent of their misunderstanding: Many believed that AI was an omniscient knowledge-producing machine connected to the internet.

After I gave a brief presentation on AI, they expressed surprise that large language models are based on prediction rather than direct knowledge. Their curiosity was piqued, and they wanted to learn how to use AI effectively.

After they drafted their speeches without AI, I asked them to use ChatGPT to proofread their drafts and then report back to me. Again, they were surprised this time about how much ChatGPT could improve their writing. I was happy (even proud) to see they were also critical of the output, with comments such as It didnt sound like me or It made up parts of the story.

Was the activity perfect? Of course not. Prompting was challenging. I noticed a clear correlation between literacy levels and the quality of their prompts.

Students who struggled with college-level writing couldnt go beyond prompts such as Make it sound smoother. Nonetheless, this basic activity was enough to spark curiosity and critical thinking about AI.

Individual activities like these are great, but without institutional support and guidance, efforts toward fostering AI literacy will fall short.

The provost of my college established an AI committee to develop college guidelines. It included professors from a wide range of disciplines (myself included), other staff members and, importantly, students.

Through multiple meetings, we brainstormed the main issues that needed to be included and researched specific topics like AI literacy, data privacy and safety, AI detectors and bias.

We created a document divided into key points that everyone could understand. The draft document was then circulated among faculty and other committees for feedback.

Initially, we were concerned that circulating the guidelines among too many stakeholders might complicate the process, but this step proved crucial. Feedback from professors in areas such as history and philosophy strengthened the guidelines, adding valuable perspectives. This collaborative approach also helped increase institutional buy-in, as everyones contribution was valued.

Related: A new partnership paves the way for greater use of AI in higher ed

Underfunded public institutions like mine face significant challenges integrating AI into education. While AI offers incredible opportunities for educators, realizing these opportunities requires substantial institutional investment.

Asking adjuncts in my department, who are grossly underpaid, to find time to learn how to use AI and incorporate it into their classes seems unethical. Yet, incorporating AI into our knowledge production activities can significantly boost student outcomes.

If this happens only at wealthy institutions, we will widen academic performance gaps.

Furthermore, if only students at wealthy institutions and companies get to use AI, the bias inherent in these large language models will continue to grow.

If we want our classes to ensure equitable educational opportunities for all students, minority-serving institutions cannot fall behind in AI adoption.

Cristina Lozano Argelles is an assistant professor of interpreting and bilingualism at John Jay College, part of the City University of New York, where she researches the cognitive and social dimensions of language learning.

This story about AI literacy was produced by The Hechinger Report, a nonprofit, independent news organization focused on inequality and innovation in education. Sign up for Hechingers newsletter.

The Hechinger Report provides in-depth, fact-based, unbiased reporting on education that is free to all readers. But that doesn't mean it's free to produce. Our work keeps educators and the public informed about pressing issues at schools and on campuses throughout the country. We tell the whole story, even when the details are inconvenient. Help us keep doing that.

Join us today.

Original post:

TEACHER VOICE: My students are afraid of AI - The Hechinger Report

Hebbia, a startup using generative AI to search large documents and return answers, has raised a nearly $100 million Series B led by Andreessen Horowitz, according to three people with knowledge of the matter.

The round valued the company between $700 million and $800 million, although TechCrunch couldnt verify whether that valuation is pre- or post-money. (One possible scenario is $700 million pre/$800 million post.) Hebbia disclosed in an SEC filing in May that it had by then raised $93 million out of a hoped-for $100 million, but we understand from two of the people that the round hit a near $100 million mark and has closed.

Hebbia and Andreessen Horowitz didnt respond to a request for comment.

Hebbia was founded in 2020 by George Sivulka, who launched the company while working on his PhD in electrical engineering at Stanford. Sivulka was inspired by his friends working in the financial industry who told him that part of their long work weeks was spent searching for information in SEC filings and other dense documents. Sivulka thought that AI could help them save hours at the office and give them more time for rest and sleep.

Hebbias AI can look over billions of documents at once, including PDFs, PowerPoints, spreadsheets and transcripts and return specific answers, the company says.

The startup sells primarily to financial service firms, including hedge funds and investment banks. But its product could also be used by law firms and other professional domains.

The latest funding brings Hebbias total capital to over $120 million. The company raised its $30 million Series A in September 2022 led by Index Ventures with participation from Radical Ventures.

The companys product is similar to Glean, whose software can fetch information in plain English from various business applications. In February, Glean raised a $200 million Series D at a valuation of $2.2 billion, led by Kleiner Perkins and Lightspeed.

Read more here:

Hebbia raises nearly $100M Series B for AI-powered document search led by Andreessen Horowitz - TechCrunch

One of the key ingredients that made ChatGPT a ripsnorting success was an army of human trainers who gave the artificial intelligence model behind the bot guidance on what constitutes good and bad outputs. OpenAI now says that adding even more AI into the mixto help assist human trainerscould help make AI helpers smarter and more reliable.

In developing ChatGPT, OpenAI pioneered the use of reinforcement learning with human feedback, or RLHF. This technique uses input from human testers to fine-tune an AI model so that its output is judged to be more coherent, less objectionable, and more accurate. The ratings the trainers give feed into an algorithm that drives the models behavior. The technique has proven crucial both to making chatbots more reliable and useful and preventing them from misbehaving.

RLHF does work very well, but it has some key limitations, says Nat McAleese, a researcher at OpenAI involved with the new work. For one thing, human feedback can be inconsistent. For another it can be difficult for even skilled humans to rate extremely complex outputs, such as sophisticated software code. The process can also optimize a model to produce output that seems convincing rather than actually being accurate.

OpenAI developed a new model by fine-tuning its most powerful offering, GPT-4, to assist human trainers tasked with assessing code. The company found that the new model, dubbed CriticGPT, could catch bugs that humans missed, and that human judges found its critiques of code to be better 63 percent of the time. OpenAI will look at extending the approach to areas beyond code in the future.

Were starting work to integrate this technique into our RLHF chat stack, McAleese says. He notes that the approach is imperfect, since CriticGPT can also make mistakes by hallucinating, but he adds that the technique could help make OpenAIs models as well as tools like ChatGPT more accurate by reducing errors in human training. He adds that it might also prove crucial in helping AI models become much smarter, because it may allow humans to help train an AI that exceeds their own abilities. And as models continue to get better and better, we suspect that people will need more help, McAleese says.

The new technique is one of many now being developed to improve large language models and squeeze more abilities out of them. It is also part of an effort to ensure that AI behaves in acceptable ways even as it becomes more capable.

Earlier this month, Anthropic, a rival to OpenAI founded by ex-OpenAI employees, announced a more capable version of its own chatbot, called Claude, thanks to improvements in the models training regimen and the data it is fed. Anthropic and OpenAI have both also recently touted new ways of inspecting AI models to understand how they arrive at their output in order to better prevent unwanted behavior such as deception.

The new technique might help OpenAI train increasingly powerful AI models while ensuring their output is more trustworthy and aligned with human values, especially if the company successfully deploys it in more areas than code. OpenAI has said that it is training its next major AI model, and the company is evidently keen to show that it is serious about ensuring that it behaves. This follows the dissolvement of a prominent team dedicated to assessing the long-term risks posed by AI. The team was co-led by Ilya Sutskever, a cofounder of the company and former board member who briefly pushed CEO Sam Altman out of the company before recanting and helping him regain control. Several members of that team have since criticized the company for moving riskily as it rushes to develop and commercialize powerful AI algorithms.

Dylan Hadfield-Menell, a professor at MIT who researches ways to align AI, says the idea of having AI models help train more powerful ones has been kicking around for a while. This is a pretty natural development, he says.

Hadfield-Menell notes that the researchers who originally developed techniques used for RLHF discussed related ideas several years ago. He says it remains to be seen how generally applicable and powerful it is. It might lead to big jumps in individual capabilities, and it might be a stepping stone towards sort of more effective feedback in the long run, he says.

Go here to see the original:

OpenAI Wants AI to Help Humans Train AI - WIRED