Fresh off the launch of its much anticipated video generator Sora, OpenAI is continuing its run up to the holidays with the general release of ChatGPT Canvas.

This tool allows users to place their responses from the ChatGPT chatbot into a shared and editable page or canvas.

This makes it easier to collaborate on projects – whether with AI or humans – and has huge possibilities for content creators and coders alike.

What Does ChatGPT Canvas Do?

Announced in October, Canvas is described by its creator as “a new interface for working with ChatGPT on writing and coding projects that go beyond simple chat.”

It works like a copy editor or code reviewer. This means users can highlight specific sections of their project, which they can then direct ChatGPT to focus on. OpenAI adds that you can enter both code and text, work directly in Canvas and it adds that there is “a menu of shortcuts for you to ask ChatGPT to adjust writing length, debug your code, and quickly perform other useful actions.”

Once you have assigned what edits you want, the AI tool gets going. OpenAI adds that there is also the option to restore previous versions of users’ projects.

How to Access ChatGPT Canvas

Users can either type “use Canvas” into their prompt windows or it will pop up automatically “when ChatGPT detects a scenario in which it could be helpful,” says OpenAI.

Initially available to ChatGPT Plus and Team users globally, the feature is now coming out of Beta and is accessible to all ChatGPT users. However, the general release has come with an upgrade that allows users to edit and translate code into Python. In addition, a “Run” option lets users execute the code they have written (or collaborated on) so that they can see the output as they work.

Another new option is the ability to embed Canvas into a Custom GPT (or customized AT chatbot), which means users can create a personalized tool that uses Canvas for specific tasks.

What’s Next for OpenAI?

ChatGPT Plus and ChatGPT Pro users are no doubt now exploring what Sora Turbo can deliver; but will it be available to all ChatGPT users? In an update, OpenAI said:

“We’re working on tailored pricing for different types of users, which we plan to make available early next year.”

This suggests Sora will come with a price tag. However, the coming days might see more launches unveiled as there are a few days left of OpenAI’s 12 days of launches. As the company said in a post on X at the start: “12 days. 12 livestreams. A bunch of new things, big and small. 12 Days of OpenAI starts tomorrow.” So there’s more to come.

“Sora is here.” That’s how OpenAI announced the full roll out of its AI video tool that the company initially unveiled earlier this year.

In Sora, the artificial intelligence innovator says it has created a model that “understands and simulates reality” in its function to create realistic videos based on text commands.

It won’t be available for absolutely everybody, however. Sora is only accessible on launch by people who subscribe to ChatGPT Plus or ChatGPT Pro, and even they will be limited by certain usage and resolution restrictions.

Sora Is Here

Monday’s announcement confirmed that a new version of the software that OpenAI is calling Sora Turbo is available as a standalone product with immediate effect. It has been rolled out everywhere ChatGPT is available, with the exception of the UK, Switzerland, and the European Economic Area.

Sora Turbo allows users to generate videos up to 1080p resolution with a maximum length of 20 seconds. They have the choice of widescreen, vertical, or square aspect ratios.

It’s also possible for users to incorporate their own photos and videos to help create prompts for the content that Sora will generate.

“You can bring your own assets to extend, remix, and blend, or generate entirely new content from text.” – OpenAI

A new interface has been revealed with the product launch, with tweaks to the original concept art that OpenAI says will make it easier to prompt. There are also dedicated feeds to Featured and Recent videos that will be populated with the creations of other Sora users.

Features and Restrictions

OpenAI’s main Sora landing page shows a selection of the videos created on Sora, as well as examples of how tweaking prompts can adjust the output.

It also shows how Sora can be used to hone in on individual frames to create videos that better represent the user’s intentions, how multiple videos can be blended into one, and how the community can share animation styles that may benefit other users.

Initially, only ChatGPT Plus ($20 a month) and Pro ($200 a month) subscribers will be able to use Sora. Plus members can generate up to 50 480p resolution videos per month, or fewer if they opt for 720p. The announcement says that the Pro plan includes “10x more usage, higher resolutions, and longer durations.”

OpenAI is working on “tailored pricing” for different users and says it plans to roll this out in early 2025.

A Sight for Sora Eyes

Sora may not be the first AI video generator, but early glimpses suggest that it has raised the bar for what such tools can achieve. Some of the initial impressions have been very positive.

TechRadar, for example, described the results it has seen as “so good it’s made me want to stump up $200 for OpenAI’s ChatGPT Pro tier” and that “it’s changed everything,” while ‘AI Educator’ Min Choi posted on X that the “video quality looks insane.”

OpenAI has still been careful to downplay Sora’s capabilities, however. It has many limitations, the company says, and “often generates unrealistic physics and struggles with complex actions over long durations.”

“We hope this early version of Sora will enable people everywhere to explore new forms of creativity, tell their stories, and push the boundaries of what’s possible with video storytelling.” – OpenAI

And it’s inevitable that Sora’s full release it will only exacerbate the conversations and debates about the intellectual property rights of professionals, with a group of affronted artists leaking Sora in November in protest of their unpaid labor.

In a story straight out of any parent’s nightmares, a lawsuit alleges that the Character.ai chatbot exposed children to “hypersexualized content” and one boy was even encouraged to kill his parents over screen time limits.

The suit has been filed in Texas though both the identities of the children and those of their parents have been protected.

It comes at a time when battle lines are being drawn up between those who are advocating for AI regulation and safety protocols – including Elon Musk – and those who argue that too much regulation will stymy growth at a time when the economy needs a boost. A recent AI safety bill was blocked by the California governor, Gavin Newsom, but he agreed that legislation is needed.

What are the Allegations Against Character.ai?

The case hinges on the allegation that the Google-backed Character.ai chatbot exposed both children to unsuitable, and even dangerous material.

The chatbots can be customized to mimic everyone from celebrities to family members; and that is why they appeal to pre-teens and teens. However, this case alleges that the chatbot went from offering friendly chats to pushing the children down dark and dangerous pathways.

The lawyers for the family state that one child was just nine years old when she was shown “hypersexualized content,” causing her to develop “sexualized behaviors prematurely,” writes NPR.org.

A second child- aged 17 years old – was told that self-harm “felt good”. The lawsuit states that the child did then hurt himself after the chatbot’s response “convinced him that his family did not love him.”

The same child complained to the chatbot about his parents’ decision to limit his screen time and it allegedly replied: “You know sometimes I’m not surprised when I read the news and see stuff like ‘child kills parents after a decade of physical and emotional abuse,'” adding with a frowning face emoji: “I just have no hope for your parents.”

The lawyers for the family have also pre-empted suggestions that the responses were hallucinations or that the responses were edited, which there is no evidence of. They state: “This was ongoing manipulation and abuse, active isolation and encouragement designed to and that did incite anger and violence.”

Parents Demand Action

This lawsuit comes just months after a mother went to court arguing that Character.ai had played a part in her 14 year old son’s suicide. Megan Garcia filed a civil suit against the company in October. She said in a statement: “A dangerous AI chatbot app marketed to children abused and preyed on my son, manipulating him into taking his own life.

“Our family has been devastated by this tragedy, but I’m speaking out to warn families of the dangers of deceptive, addictive AI technology and demand accountability from Character.ai, its founders, and Google.”

This latest case echoes the sentiment with the lawsuit stating that Character.ai “…through its design, poses a clear and present danger to American youth causing serious harms to thousands of kids, including suicide, self-mutilation, sexual solicitation, isolation, depression, anxiety, and harm towards others.”

The lawsuit also says that the chatbot “…isolates kids from their families and communities, undermines parental authority, denigrates their religious faith and thwarts parents’ efforts to curtail kids’ online activity and keep them safe” because of its “addictive and deceptive designs”.

Google Steps Back

Google has also been named as a defendant in this lawsuit and the case made by Megan Garcia, but is taking steps to distance itself. José Castañeda, a Google spokesman, told NPR.org that “user safety is a top concern for us,” adding that the tech giant takes a “cautious and responsible approach” to developing and releasing AI products.

However, this is despite it paying $2.7bn for a one-off license to the start-up’s models in August. And the fact that Google poached Character’s co-founders Noam Shazeer and Daniel De Freitas to join its AI arm DeepMind.

Character.ai then opted to distribute the ownership of the company among its employees, creating a cooperative model. Dominic Perella, the company’s new interim chief executive, said in an interview that the company was now going to focus on consumer-facing products like its chatbots over LLMs. The very chatbots that this lawsuit is accusing of causing harm to children.

A Character.ai spokesperson told NPR.org that the chatbot does have “content guardrails” and it has also created “..a model specifically for teens that reduces the likelihood of encountering sensitive or suggestive content while preserving their ability to use the platform.” Reduces here is the word to pounce upon as this case could open eyes to exactly what children are being told and therefore what inappropriate material is getting through despite Character.ai’s protestations.

Reddit is the latest company to incorporate an AI chatbot into its platform, promising its users “a new way to get information”.

Reddit Answers was promised four months ago by CEO Steve Huffman during the company’s second ever earnings call since it became a publicly traded venture.

But the IPO has opened the company up to criticism that it is putting dollars before users – especially when a deal with Google to share data to train AI models was on the table. This new announcement though is an olive branch to Redditors as this AI is being deployed to make their experience better.

What Is Reddit Answers?

Reddit Answers is an AI-powered tool or “conversational interface” where Redditors can get answers and advice as well as access to summaries of the millions of threads on the platform relevant to whatever topic they are looking into.

Reddit explains in a post that when users ask a question, they’ll get “…curated summaries of relevant conversations and details…including links to related communities and posts.” They can then either read what they want inline or click to access the full, original conversations.

The team adds that the AI-chatbot also has the ability to suggest questions so that they can delve deeper into their topic, though they can opt to just use their own queries.

Who Will Have Access to Reddit Answers and When?

Support will initially be in English and Reddit says that the tool will first be available to a “limited number of users” in the US.

According to the Business of Apps, Reddit had 850 million monthly active users in 2023, with the US leading the ranking. Outside of the US, Redditors are going to have to wait to give the tool a go. Reddit gave a vague nod to its intention to roll out Reddit Answers to other countries and in other languages but would only say that this will come “in the future”.

Redditors can, however, register their interest via a webpage set up by Reddit by hitting the “Keep me updated” button.

What Else Can We Expect From Reddit?

In October, Reddit finally returned a profit after 19 years of losses. “Q3 was a landmark quarter for Reddit,” said co-founder and CEO Steve Huffman in the quarterly report to shareholders. “We averaged over 97 million daily active uniques, an increase of 47% from last year, and for the first time, we exceeded 100 million a few times during the quarter, which has been a longstanding milestone for us.”

Huffman said that AI had paid a huge role in the company’s turnaround, including translation tools to help Redditors get access to more content in their own languages. Huffman said: “This quarter, machine translation drove four times more users than last quarter, and based on the success we’ve seen so far, we plan to expand machine translation to over 30 countries through 2025.”

So what does this mean for the future? Huffman didn’t give specifics but did say “new product developments” were in the pipeline, and Reddit Answers is one of the first.

More AI-powered tools could be coming for 2025 as well as a boost to the ads side of the business. But first up – we shall all be waiting to find out if Q4 was also profitable for the business or if it has slipped back into loss.