22 Feb AI is entertaining but truth & success rely on quality data
ChatGPT write me a poem… The buzz on chatbots has reached a fever pitch.
Making AI conversational chatbots like Chat GPT sound human and confident is proving easier than making them accurate, as public launch stumbles of Bard (Alphabet) and Bing (Microsoft) have shown.
When chatbots crowdsource the internet to create an AI narrative, inevitably you face ‘truth issues’ and biases inherent to the internet. Not surprisingly, two-thirds of executives in the new LxT Path to AI Maturity 2023 study say their need for high-quality data is increasing as hungry machine learning models continue to gobble up training data to build and improve.
Interestingly, AI cost is not a dominant driver for organizations deploying AI with half of the $500M+ revenue organizations surveyed having AI budgets over $75 million. Instead, AI strategies are being driven by the need for innovation and agility, anticipating customer needs and quickly bringing products to market.
Conversational AI solutions like chatbots are reportedly generating the highest returns (ROI), but 46% of AI projects failing to meet their goals, paying attention to a quality data strategy supporting AI has never been more critical.
LXT commissioned Reputation Leaders in October 2022 to survey 315 senior decision-makers working for US organizations who had verified AI experience on their views and predictions for AI and data strategies. The Path to AI Maturity 2023 is downloadable from the LxT website
This was first published on LinkedIn on February 17, 2023