# Crafting Realistic Fake News with AI: A Closer Look at NASA's Mars Breakthrough
Written on
Chapter 1: Analyzing AI-Generated News
Imagine a scenario where a news anchor announces a significant advancement by NASA in Mars exploration. Surprisingly, this news is fabricated and crafted entirely through artificial intelligence. The question arises: how many different AIs were utilized to produce such a believable video?
To find out, let’s delve into the AIs involved in creating this strikingly authentic fake news.
Section 1.1: The Power of GPT-3
The journey begins with OpenAI's GPT-3. By inputting a brief tagline, such as "Breakthrough in Mars project," we can obtain a detailed narrative crafted by the AI. Utilizing the GPT-3 Playground, we generate text that serves as the foundation for our fabricated news report. The results are impressive, providing a compelling storyline to work from.
Section 1.2: Visuals from DALL-E-2
Next, we turn to DALL-E-2 to create a stunning background image that aligns with our headline. By entering the same tagline, the AI produces four striking images, from which I select the most impactful. DALL-E-2 serves as our visual artist, enhancing the overall presentation of our fabricated news.
Subsection 1.2.1: Creating the Avatar with Synthesia
Now, we introduce Synthesia to add a dynamic element to our video. This AI enables us to generate a moving avatar along with a synthetic voice. After choosing the avatar named Rosie, we upload the background image from DALL-E-2.
Next, we transfer the text generated by GPT-3 into Synthesia. For the voice, I select the "English (US) — Professional" option, ensuring a polished delivery.
Finally, we generate the video, which showcases a fully realized AI-produced news segment.
Chapter 2: Summary of AI Integration
In summary, we have effectively merged three distinct AIs to fabricate a convincingly realistic news piece about a NASA breakthrough related to Mars exploration. Depending on how one counts, it could even be considered four AIs, as Synthesia manages to create both an avatar and a synthetic voice.
The AIs employed in this project include:
- OpenAI Playground: for generating text based on a brief tagline.
- DALL-E-2: for producing an image that corresponds with the tagline.
- Synthesia: for creating both the avatar and the synthetic voice.
If you're interested in following my journey and supporting my work, feel free to sign up on Medium. Should you have any questions or require assistance, don't hesitate to leave a comment—I’ll be more than happy to help!