{"id":62563,"date":"2023-02-07T09:22:30","date_gmt":"2023-02-07T17:22:30","guid":{"rendered":"https:\/\/www.salesforce.com\/?p=62563"},"modified":"2023-08-30T14:51:39","modified_gmt":"2023-08-30T21:51:39","slug":"building-generative-ai-we-can-trust","status":"publish","type":"post","link":"https:\/\/www.salesforce.com\/news\/stories\/building-generative-ai-we-can-trust\/","title":{"rendered":"Building Generative AI We Can Trust"},"content":{"rendered":"\n

Editor\u2019s Note: AI Cloud, Einstein GPT, and other cloud GPT products are now Einstein. For the latest on Salesforce Einstein, go <\/em>here<\/em><\/a>.<\/em><\/p>\n\n\n\n

2022 will be remembered as an inflection point for generative AI \u2014 AI that doesn\u2019t just classify or predict, but creates<\/em> content of its own, be it text, imagery, video, or even executable code. And, it does so with a human-like command of language.\u00a0<\/p>\n\n\n\n

It was the year that we saw large foundation models (a deep learning algorithm that has been pre-trained with large data sets with a wide variety of data, which can transfer knowledge from one task to another) like ChatGPT, StableDiffusion, and Midjourney drive widespread attention by delivering capabilities that would have seemed like science fiction only years ago. <\/p>\n\n\n\n

But despite their flashy output, these models\u2019 most striking feature might be their sheer flexibility. Unlike even the most advanced machine learning models of the last decade, foundation models appear capable of producing truly incredible content and solving a staggering range of problems \u2014 from writing poetry and explaining physics, to solving riddles and painting pictures. And with additional training and human guidance, their applicability can be extended even further.<\/p>\n\n\n\n

What do foundation models mean for the future of work?<\/h2>\n\n\n\n

As a technologist and researcher, I\u2019ve long believed that AI won\u2019t just enhance the way we live, but transform it fundamentally. In particular, the more I explore Conversational AI<\/a>, the more convinced I am that it will increasingly dissolve the correlation between a technology\u2019s power and the expertise required to harness it. AI is placing tools of unprecedented power, flexibility, and even personalization into everyone\u2019s hands, requiring little more than natural language to operate. They\u2019ll assist us in many parts of our lives, taking on the role of superpowered collaborators.<\/p>\n\n\n\n

For engineers, marketers, sales reps, and customer support specialists, the role of AI in day-to-day work will only grow in the coming years. At Salesforce, we\u2019ve spent years embedding state-of-the-art AI within business applications spanning sales, service, marketing, and commerce, and today, our Customer 360 platform<\/a> is generating more than 200 billion AI-powered predictions per day.<\/p>\n\n\n\n

\n

But is the revolutionary promise of generative AI truly at hand, or somewhere on the horizon? It\u2019s a more complex topic than the press coverage might suggest. <\/p>\n<\/blockquote>\n\n\n\n

My role<\/a> as Executive Vice President and Chief Scientist of Salesforce Research<\/a> has given me a unique perspective on this. We operate within the world\u2019s biggest companies, reaching billions of people in one form or another, and serve industries that touch every facet of society. That means everything we put in the hands of our customers has to offer mission-critical reliability<\/em> as well \u2014 the kind that engenders lasting trust. <\/p>\n\n\n\n

And while no one denies the power of generative AI as a whole, our ability to trust it is another matter entirely.<\/p>\n\n\n\n

The dual nature of generative AI: new capabilities paired with new risks<\/h2>\n\n\n\n

Generative AI promises an entirely new way to interact with machine intelligence, but it introduces what might be an entirely new kind of failure as well \u2014 confident<\/em> failure. The poised, often professional tone these models exude when answering questions and fulfilling prompts make their hits all the more impressive, but it makes their misses downright dangerous. Even experts are routinely caught off guard by their powers of persuasion. <\/p>\n\n\n\n

For example, in December 2022, researchers at the University of Chicago and Northwestern University<\/a> used ChatGPT to generate abstracts based on titles taken from real articles in five medical journals. Then, when given a mix of original and fictitious abstracts in a blinded review, expert reviewers misidentified 32% of those generated by ChatGPT as originals and incorrectly identified 14% of the originals as being generated by ChatGPT. <\/p>\n\n\n\n

This is a complex challenge, but it\u2019s an easy one to understand. In a forthcoming paper<\/a>, a team of cognitive scientists from MIT, UCLA, and UT Austin analyzed the models that power generative AI in terms analogous to the human brain. Their conclusion was twofold: foundation models like the large language models (LLMs) trained on textual data have mastered one aspect of our capacity for language \u2014 formal linguistic competence<\/em>, which describes the complex but superficial ability to follow rules like grammar, conjugation, and word choice \u2014 while barely scratching the surface of functional linguistic competence<\/em>, which refers to the use of language to express non<\/em>-linguistic skills like knowledge of history, common sense, and reasoning ability. <\/p>\n\n\n\n

\n

Simply put, LLMs are masters of language, but only<\/em> language<\/a>. Their ability to craft a perfectly written paragraph is, unsettlingly, entirely divorced from what the paragraph\u2019s component sentences actually mean.<\/p>\n<\/blockquote>\n\n\n\n

That\u2019s a dangerous imbalance and one that simply must be rectified before these tools can play a mission-critical role in the real world.<\/p>\n\n\n\n

Generative AI use cases vary \u2014 and so does the challenge of trust<\/h2>\n\n\n\n

But given the sheer scope of this technology\u2019s potential across the enterprise, I\u2019m certain this is a challenge worth confronting. After all, generative AI can help create complex content like artwork, prose, and even code; summarize information from a range of sources and deliver it in abbreviated form or an interactive question-and-answer session; facilitate lifelike conversational search and information retrieval; help better understand analytics, and more. These are revolutionary possibilities, and they\u2019re well worth exploring.<\/p>\n\n\n\n

Human intervention remains necessary in each of these use cases, but to varying degrees and at varying points in the process. For instance, a human editor may fact-check and refine AI-generated marketing copy before it\u2019s put to use. In other cases, such as a non-designer relying on AI to design an ad banner or email layout, the feedback may be binary \u2014 accepted or not. In more advanced applications, AI can play a collaborative role, helping experts automate the lower-level tasks of a project while their creativity is focused on the more sophisticated challenges. <\/p>\n\n\n\n

Consider, for example, an IT administrator partnering with a code-generation model in the creation of a custom application. The AI can flesh out routine components and subsystems in response to simple prompts \u2014 for example, \u201corganize the log files in directories named after each network, and automatically delete existing files over six months old\u201d \u2014 while the human focuses on developing the kind of novel, problem-solving logic that delivers unique value.<\/p>\n\n\n\n