BENGALURU: Last month, South Korean consumer electronics giant Samsung Electronics discovered a security risk. Some of the company’s employees have reportedly uploaded sensitive code to his ChatGPT, an artificial intelligence (AI) chatbot developed by OpenAI, which interacts with users in a conversational fashion. Additionally, we provided the bot with the recorded meeting and asked it to take the minutes.
BENGALURU: Last month, South Korean consumer electronics giant Samsung Electronics discovered a security risk. Some of the company’s employees have reportedly uploaded sensitive code to his ChatGPT, an artificial intelligence (AI) chatbot developed by OpenAI, which interacts with users in a conversational fashion. Additionally, we provided the bot with the recorded meeting and asked it to take the minutes.
Samsung has since banned the use of such AI tools by its employees. Internal memos reportedly inferred that data sent to AI platforms such as ChatGPT and Google Bard would be stored on external servers. This makes data difficult to retrieve and delete, potentially compromising sensitive company information.
Hi!reading premium articles
Samsung has since banned the use of such AI tools by its employees. Internal memos reportedly inferred that data sent to AI platforms such as ChatGPT and Google Bard would be stored on external servers. This makes data difficult to retrieve and delete, potentially compromising sensitive company information.
The incident at Samsung provides a glimpse into the Pandora’s box of concerns that global companies facing the dizzying pace of generative AI can open. Since its release five months ago, ChatGPT has gained over 100 million users. Individuals (and even small businesses) around the world use it for everything from writing blogs, reviews, and resumes, to creating short films and realistic images, to generating software his code, to analyzing broad economic trends. I am using this tool. All without human intervention.
Growing interest makes it impossible to ignore generative AI, but big companies are probably right to proceed cautiously.
Chatbots like ChatGPT are more knowledgeable than most humans because they are trained on billions of words from sources such as the internet, books, and multiple online sources such as Common Crawl and Wikipedia. However, they are not necessarily intelligent. A bot might be able to connect the dots, but it doesn’t always understand what the bot spits out.
I have other concerns. The US Federal Trade Commission (FTC) has warned that these AI tools can be used by fraudsters to replicate the voices of relatives in just short audio clips and wreak havoc. For example, on April 29th, CNN Arizona woman Jennifer DeStefano reportedly believes she is the victim of a de facto kidnapping scam in which someone duplicates her daughter’s voice and demands a ransom. Similarly, scammers can use these tools to impersonate siblings, request OTPs (one-time passwords), and siphon money out of bank accounts.
A number of prominent technology leaders, including Elon Musk, Joshua Bengio, and Stuart Russell, called for a six-month moratorium on training systems “more powerful than GPT-4,” prompting the world to It should only be developed if it believes it is possible, he argued. So did Jeffrey Hinton, known for his deep learning research that powers today’s generative AI tools, recently quit Google to “speak freely about the risks of AI.”
Given this apparent current danger, midsize and large businesses, especially those in the Banking, Financial Services and Insurance (BFSI) and Healthcare sectors, are proceeding with due caution. Large financial institutions such as Citigroup, Bank of America, Deutsche Bank, Goldman Sachs, Wells Fargo and JPMorgan Chase have been criticized for exposing sensitive information while using ChatGPT technology. , has already placed restrictions on the use of ChatGPT by its employees.
The chief information officer (CIO) of a large Indian multinational bank, speaking on condition of anonymity, said: mint: “ChatGPT clearly has potential, but it’s a bit premature for the highly regulated BFSI sector. No. I would rather wait until the technology matures and there are guardrails around it.”
A use case emerges
Nonetheless, corporate boards have been flooded with conversations about generative AI, fueled by discussions about the release of ChatGPT and its potential use cases, according to the latest report, and between January and March this year. 17% of CEOs discussed this AI in the quarter. The ‘What the CEO Told’ report was produced by He IoT Analytics, a Germany-based provider of market insights and strategic business intelligence.
Fintech company Paytm made eight mentions of AI and AGI (artificial general intelligence) during its May 11 earnings call. Experts believe that in the not-too-distant future, AGI machines will be able to understand the world as humans do, and in many cases surpass human intelligence.
Microsoft, Google, International Business Machines (IBM), and Nvidia are powering generative AI platforms that businesses can use with reduced data and security concerns. For example, Microsoft has already started giving enterprise users “the tools they need to build applications powered by ChatGPT.”
OpenAI is also working on ‘ChatGPT Business’, claiming that it ‘by default does not use end-user data for model training’. Nvidia offers a cloud service (NeMo) that integrates generative AI capabilities into enterprise applications Amazon Inc. has its own generative AI platform called Bedrock, while IBM offers WatsonX and has an open Source Generation We partner with Hugging Face, an AI company. HuggingChat competes with ChatGPT.
Data security skepticism doesn’t stop enterprise use cases from budding. Travel and vacation price aggregators such as Expedia have already started using ChatGPT to provide the best flights and help travelers plan their trips and vacations.
Shopify Magic, an AI product from e-commerce platform Shopify, generates product descriptions from a list of keywords or product descriptors in tone chosen by the merchant. Meanwhile, retail giant Carrefour is experimenting with using ChatGPT to create videos that answer customer questions such as “how to”. Eat healthier for less. ”
Generative AI also has use cases in the human resources (HR) world. Tasks such as onboarding, training, performance management, employee inquiries and complaints can be automated using ChatGPT. In finance, AI can help with compliance, credit risk management, investment research, and legal document processing.
indian play
In India, the Mahindra Group is considering several use cases for its business units. “Generative AI is evolving at a fast pace and exploring the right use cases for our business is something we are very excited about. We are focused on making the best use of this technology for our business,” said Rucha Nanavati, Chief Information Officer of Mahindra Group.
Walmart-owned Flipkart also has a lot of potential for generative AI to address one of the core problems that every e-commerce platform is trying to solve: connecting the products consumers want to buy. I believe there is
“Gen AI will enable us to build more conversations, human-like agents and assistants to handle users throughout the discovery, purchase and post-purchase customer service journey,” said Flipkart’s best products and technology. Chief Jeyandran Venugopal explained. “AI helps us build high-quality content (pictures and descriptions) for our product catalogs and merchandising and advertising campaigns. AI summarizes product descriptions and user reviews to provide It helps reduce cognitive load,” he added.
Sanjay Mohan, group CTO of MakeMyTrip, is currently using generative AI for proof-of-concept (PoC) work. According to Mohan, Large Language Models (LLM) are very good at summarizing things very smartly and clearly. “For reviews, if someone says ‘outstanding’ and the other says ‘great’, you know they are both the same. So that summary is what we can use. ”
Audio platform Pocket FM uses generative AI to automate the creation of long-tail content (using specific keywords such as ‘size 7 hiking boots for men’ instead of just ‘hiking boots’), trailers and promotions. , to provide personalized recommendations. According to his co-founder and CTO Prateek Dixit, users his data is important. He added that adopting the technology has reduced Pocket FM’s translation time by more than 40%.
Mumbai-based startup Haptik uses generative AI to make bot conversations less robotic and more free-flowing, said co-founder and CEO Akrit Vaish. Haptik also hopes to use ChatGPT to “create content and add countless variations of bot responses.” His company, Zoho Corp., has also started extending and integrating 13 generative AI Zoho applications powered by ChatGPT.
limit
However, integrating application programming interfaces (APIs allow applications to communicate with each other) with other departments’ business workflows presents a unique set of challenges for enterprises. As Sumanta Kar, technology partner at consulting firm EY India, points out, once you have a tool like ChatGPT in place, you need to continuously monitor, Needs retraining, fine-tuning. -to date.
According to Sanjeev Menon, co-founder and head of technology and product at E42.ai, a natural language processing-based AI platform, ChatGPT is great at generating text and answering questions, but has complex workflows. may not be very capable of automating It explains why in an enterprise environment such a model should be fine-tuned before using it.
That’s because corporate data includes structured and unstructured data such as videos, audio files, social media posts, and emails. Organizations therefore rely on specialized tools that can interact with third parties and internal systems. It also performs certain actions based on the collected data.
“There is no customer meeting today without a discussion of generative AI,” said Jaya Kishore Reddy, co-founder and CTO of conversational AI startup Yellow.ai. But he also emphasized the need for an “orchestration layer” to connect. Applying generative AI models like ChatGPT to business systems requires significant customization. “Even plug-ins (tools that help ChatGPT to access the latest information, perform calculations, use third-party services, etc.) connect to different systems and multiple workflows within the enterprise. We need to be able to,” he added.
Bharath Shankhar, vice president of engineering at conversational AI company gnani.ai, emphasized the importance of defining boundaries or scopes in specific domains to make GPT systems efficient. Shankar also stressed that a lot of effort is needed to make GPT “integrate with corporate backend systems such as ticketing tools and CRMs”, and companies confirm that there are no regulatory violations. said I need to. For example, accessing or sharing patient data could lead to his HIPAA (Health Insurance Portability and Accountability Act) violations in the healthcare sector, and furthermore, the ChatGPT bot response time is I pointed out that it might not work in a scenario like that. Real-time response without delay is required.
Since then, however, the power of generative AI models has grown rapidly. IBM predicts that self-supervised learning will soon be used in so-called “foundational models” (models that can be trained on a wide range of datasets and used for a variety of tasks). The ability to apply what you learn to specific tasks dramatically accelerates the adoption of AI in your business.
The blistering pace of self-training of these models and the imminent introduction of ChatGPT Business do not bode well for executives on the sidelines.
Prasid Banerjee and Abhijit Ahaskar contributed to the story
