✍️ ChatGPT Takes the Accounting Exam: Gets a Pass, But No A+ 👎

In today’s Tracker:

Read time: 4 mins

• 🧪 Research highlights: ChatGPT Struggles to Balance Its Books in Accounting Exams

• 🚨 Industry news: Generative AI Gives Customer Service Workers a 14% Boost

• ⚖️ Policy and regulation: AI's Liability Quandary: Is There a Simple Fix?

• 🌐 AI and society: The Doctor Will See You Now...with AI Assistance

• 🧰 Tool of the day: YOUS: Don’t Let Language Barriers Hold You Back

• 🧠 Let’s get smart with AI: Weak AI: It's Not the Size of the AI, It's How You Use It

📚 ChatGPT Struggles to Balance Its Books in Accounting Exams

Brigham Young University (BYU) researchers and 186 other universities examined OpenAI's AI chatbot, ChatGPT, to determine how it would perform on accounting tests.

They supplied the bot with 25,181 classroom accounting test questions of varying sorts, difficulty levels, and subjects.

They discovered that ChatGPT scored 47.4%, whereas the average student scored 76.7%.

The AI bot performed better in auditing and accounting information systems but suffered in managerial, financial, and tax evaluations, which could be attributed to a lack of mathematical processing capabilities.

Regardless of the findings, the researchers believe that technology will revolutionize how teachers and students learn.

📈 Generative AI Gives Customer Service Workers a 14% Boost

According to a new study by researchers at Stanford University and the Massachusetts Institute of Technology, customer service employees at a Fortune 500 software company with access to generative AI tools were 14% more productive on average than those without.

This is the first time outside a laboratory that the impact of generative AI tools on work has been measured.

The researchers monitored the performance of over 5,000 customer service representatives in the Philippines based on critical metrics.

The research suggests that the increase in the productivity of low-skilled workers may stem in part from the ability of AI tools to assimilate tacit knowledge and then disseminate it to less-skilled workers via AI-generated suggested responses.

🤷‍♀️ AI's Liability Quandary: Is There a Simple Fix?

Who should be held accountable for atrocities caused by AI is a topic of debate.

Some argue that AI manufacturers should be held liable for negative consequences caused by their AI, but this may not be feasible or cost-effective.

Based on the Coase Theorem, liability should be ascribed to the party who avoided the harm at the lowest cost.

Therefore, providers of general-purpose technologies such as email and mobile devices should not be held liable for harm caused by their use.

Similarly, it may not be practical to require that AI-generated information be accurate and innocuous, given that current AI models can generate vast content in numerous languages.

A positive attitude toward AI and the adoption of beneficial adjustments voluntarily may be a more effective strategy for ensuring AI safety.

👋 The Doctor Will See You Now...with AI Assistance

Oh, dear patient, do you ever feel like your healthcare providers are strangers to you, even as experts in their field? Well, imagine a world where artificial intelligence could help bridge that gap. Picture this: before meeting with a specialist, they have already read a summary of your healthcare needs created by generative AI. During your visit, a chatbot assistant based on a foundation model could help your doctor diagnose more accurately and select the best therapy for you. And afterward, a generative model could draft a clinical note in real time, leaving more time for face-to-face discussion. And that's not all! In the back office, generative models could optimize clinic scheduling and simplify billing processes.

But wait, there's more! With recent legislation giving patients digital access to medical records, a foundation model could generate tailored patient education materials and explain your care plan at an appropriate reading level. And machine learning models, reliant on large medical datasets, can be modified to create realistic clinical images from text prompts to help diagnose even the rarest diseases.

Of course, we know that factual correctness is of the utmost importance in medicine, and generative AI has struggled in the past with accuracy. But fear not, for we (humans) are working hard to improve the accuracy of medical explanations generated by these models for safe clinical use. With these advancements, patient care and engagement could reach new heights, all thanks to the power of AI.

🗨️ YOUS: Don’t Let Language Barriers Hold You Back

YOUS is an artificial intelligence (AI) platform with a built-in translator for audio/video meetings, phone calls, chats, and messaging. Its primary purpose is to eliminate language as a communication barrier and enable seamless global collaboration.

Learn more about the tool here.

👊 Weak AI: It's Not the Size of the AI, It's How You Use It

Ah, weak AI, the not-so-mighty cousin of strong AI, who dreams of being more than just a simple algorithm. Unlike its more potent relative, weak AI is incapable of human-level intelligence, consciousness, or self-awareness. It is, however, still quite valuable for specific tasks, such as playing chess, voice recognition, or facial recognition. Think of it as a loyal and obedient servant, always ready to perform the task at hand but never daring to question its purpose or existence. So, while it may not be the star of the AI show, weak AI is still a valuable member of the digital workforce.

Nobody phrases it this way, but artificial intelligence is almost a humanities discipline. It’s really an attempt to understand human intelligence and human cognition.

It is better to fail in originality than to succeed in imitation.

• 💾 Nightcafe Studio - an online platform allowing creatives to easily create, store and distribute digital media content.

• 🛠️ Altered.ai - is a cloud-based platform designed to help users build, deploy, manage, and monitor AI-driven applications.

• 🤖 Bertha.ai - an artificial intelligence (AI) platform that allows businesses to automate tasks and processes that would otherwise require human intervention.

If you liked this edition, please share it with a friend or two! You could get some exclusive content in your inbox or even some merch very soon👇🏾

And that's a wrap!

Did we miss anything? Or want to say hey? Hit reply - I'd love to hear from you! You can

• 🐦 Find us on Twitter

• 💬 Huzzah on LinkedIn

• 📱 Slide into our DMs on Instagram

And if you haven’t already, sign up to get this in your inbox daily.

Did you miss our last newsletter?