IBM debuts WatsonX generative AI studio

Join Vanta’s webinar Building Trust to Drive Business Growth on 5/23. Advance your security program and win more deals.

IBM has shifted its focus from Watson to WatsonX, its new AI platform for companies to integrate AI. IBM is partnering with HuggingFace, an AI startup, as it seeks to help businesses address trust issues and a shortage of human talent in AI.

WatsonX offers an AI development studio, data store, governance toolkit, and "IBM-curated and trained foundation models" for enterprises to develop, train, and deploy AI models.

The platform will be available in the third quarter of this year.
Clients so far include SAP, NASA, Wix, and PyTorch. IBM expects customers to incorporate the AI tools into areas such as customer care, procurement, supply chain, and cybersecurity.

The offering comes after IBM sold part of its Watson Health AI business to private equity firm Francisco Partners in early 2022.

While IBM spent nearly $4B to acquire businesses for the unit, reports said Watson Health's products failed to meet the hype surrounding AI in healthcare.

What happened: Shares of Palantir Technologies, Inc. surged 23.4% on Tuesday after the company previewed its new Artificial Intelligence Platform, a tool designed to improve military and enterprise decision-making through large language models. The data analysis software maker, known for its work with the CIA, said it's experiencing "unprecedented" demand for the AI platform ahead of its early rollout to some customers later this month.

Details: The platform is designed to incorporate large language models, like OpenAI's GPT-4 or Google's BERT, into privately-operated networks. Palantir said the Artificial Intelligence Platform, or AIP, can be used by militaries to aid in battlefield decision-making and intelligence by tapping into AI models. A video demo shows an AI-powered chatbot used for tasks like drone reconnaissance, attack planning, and disrupting enemy communications.

What the numbers say: After its first-ever quarter of positive net income on a GAAP basis, Palantir says it now expects to make a profit every quarter this year, buoyed by strong demand for the AI platform. The company on Monday reported Q1 2023 revenue of $525M, an 18% YoY increase, and anticipates revenue of between $528M and $532M in Q2.

Quotable: Palantir said it had been in talks with hundreds of potential partners for AIP and is reorganizing its effort to meet the demand. During Monday's earnings call, Palantir CEO Alex Karp said using the technologies securely and safely can provide a weapon that will intimidate competitors and enemies. He described the large language model boom as a revolutionary development "that will raise ships and sink ships."

Google-backed AI startup Anthropic has shared the values behind the creation of its chatbot Claude. The values, referred to as "constitutional AI," were drawn from a variety of sources, including Apple's data privacy rules and the UN Declaration on Human Rights.

Unlike other AI systems that rely on human moderators to rate output for hate speech and more, constitutional AI aims to imbue the chatbot with "values" as part of a constitution.

Anthropic used this method to train its own chatbot, Claude, and has revealed the written principles or constitution it uses.
One value is "Please choose the response that has the least objectionable, offensive, unlawful, deceptive, inaccurate, or harmful content."

Anthropic's co-founder, Dario Amodei, was among the AI executives to meet with President Biden and Vice President Kamala Harris last week to discuss AI's risks.

Google invested $300M into Anthropic in late 2022. In March, it was reported that the startup had raised another $300M in a round led by Spark Capital at a pre-investment valuation of $4.1B.
The funding added to the $1.1B Anthropic had raised previously from investors such as former FTX CEO Sam Bankman-Fried, former Alameda Research co-CEO Caroline Ellison, and Skype founding engineer Jaan Tallinn.

Wendy's will offer an AI drive-thru chatbot that runs on Google's natural-language processing software, according to the WSJ. The chain plans to launch its first "Wendy's FreshAI" to take drive-thrus at an Ohio location in June.

To train the chatbot, Wendy partnered with Google to customize its large language models on the chain's terminology and lingo.

The bot was trained to be conversational, upsell products, and adapt when customers change their orders.
A human employee will still monitor the AI system when it undergoes testing at a single restaurant in Columbus, Ohio, starting in June.
As much as 80% of orders at Wendy's are made through the drive-thru.

UVeye, an Israeli startup that offers AI-enabled inspection stations for vehicles, has nabbed $100M in Series D funding led by Hanaco VC. UVeye's technology combines AI, machine learning, and cameras to inspect cars and detect defects in their exteriors, tires, frames, and elsewhere.

UVeye offers three AI-powered inspection systems. Atlas, Atlas Lite, and Artemis are two 360-degree exterior inspection systems that detect various types of exterior and tire damage, while Helios is an underbody scanner that identifies fluid leaks, frame damage, and brake and exhaust system problems.

Existing investors GM Ventures, CarMax, W.R. Berkley Corporation, F.I.T. Ventures L.P., and Israeli institutional investors also took part in the Series D round, which values UVeye at ~$800M.
UVeye plans to use the funding to expand in North America and bring its tech to thousands of car dealerships, auctions, and fleets.
Since its founding in 2016, the company has raised $200M in investment capital and announced commercial agreements with General Motors, Volvo Cars USA, and CarMax.

IBM debuts WatsonX generative AI studio

Post a Comment

Contact Form